assessed through longitudinal mri and histopathology *,1,4...

Correlations of Behavioral Deficits with Brain PathologyAssessed through Longitudinal MRI and Histopathologyin the R6/1 Mouse Model of Huntington’s DiseaseIvan Rattray1,2, Edward J. Smith1,2, William R. Crum3, Thomas A. Walker2, Richard Gale2, Gillian P. Bates2,Michel Modo*,1,4

1 King’s College London, Institute of Psychiatry, Department of Neuroscience, London, United Kingdom, 2 King’s College London, Department of Medical andMolecular Genetics, London, United Kingdom, 3 King’s College London, Department of Neuroimaging, Institute of Psychiatry, London, United Kingdom,4 University of Pittsburgh, Department of Radiology, McGowan Institute for Regenerative Medicine, Pittsburgh, Pennsylvania, United States of America

Abstract

Huntington’s disease (HD) is caused by the expansion of a CAG repeat in the huntingtin (HTT) gene. The R6 mousemodels of HD express a mutant version of exon 1 HTT and typically develop motor and cognitive impairments, awidespread huntingtin (HTT) aggregate pathology and brain atrophy. Unlike the more commonly used R6/2 mouseline, R6/1 mice have fewer CAG repeats and, subsequently, a less rapid pathological decline. Compared to the R6/2line, fewer descriptions of the progressive pathologies exhibited by R6/1 mice exist. The association between themolecular and cellular neuropathology with brain atrophy, and with the development of behavioral phenotypesremains poorly understood in many models of HD. In attempt to link these factors in the R6/1 mouse line, we haveperformed detailed assessments of behavior and of regional brain abnormalities determined through longitudinal, invivo magnetic resonance imaging (MRI), as well as an end-stage, ex vivo MRI study and histological assessment.We found progressive decline in both motor and non-motor related behavioral tasks in R6/1 mice, first evident at 11weeks of age. Regional brain volumes were generally unaffected at 9 weeks, but by 17 weeks there was significantgrey matter atrophy. This age-related brain volume loss was validated using a more precise, semi-automated TensorBased morphometry assessment. As well as these clear progressive phenotypes, mutant HTT (mHTT) protein, thehallmark of HD molecular pathology, was widely distributed throughout the R6/1 brain and was accompanied byneuronal loss. Despite these seemingly concomitant, robust pathological phenotypes, there appeared to be littlecorrelation between the three main outcome measures: behavioral performance, MRI-detected brain atrophy andhistopathology. In conclusion, R6/1 mice exhibit many features of HD, but the underlying mechanisms driving theseclear behavioral disturbances and the brain volume loss, still remain unclear.

Citation: Rattray I, Smith EJ, Crum WR, Walker TA, Gale R, et al. (2013) Correlations of Behavioral Deficits with Brain Pathology Assessed throughLongitudinal MRI and Histopathology in the R6/1 Mouse Model of Huntington’s Disease. PLoS ONE 8(12): e84726. doi:10.1371/journal.pone.0084726

Editor: David R Borchelt, University of Florida, United States of America

Received August 14, 2013; Accepted November 18, 2013; Published December 19, 2013

Copyright: © 2013 Rattray et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was supported by grants from the Medical Research Council (G0800846) and the CHDI Foundation. The funders had no role in studydesign, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

* E-mail: [email protected]

Introduction

Huntington’s disease (HD), a progressive, autosomaldominant neurological disorder, is caused by a CAG/polyglutamine repeat expansion [1]. This unstable elongationleads to the eventual aggregation of mutant huntingtin (mHTT),resulting in neurodegeneration and ultimately death. Evenduring the prodromal phase, subtle brain changes have beenshown to be associated with disease burden [2,3,4]. Still,despite the genetic cause for HD being identified, the causalcascade linking the activity of mHTT with the development ofclinical signs remains poorly understood. The varying nature ofHD, as well as the scarcity of affected tissue during the early

stages of disease, impede efforts to link molecular pathology toboth brain atrophy and motor/behavioral dysfunction.

To overcome such issues and improve our understanding ofthe causes of HD, a variety of rodent models have beendeveloped. Transgenic mice include the R6 [5] and N171Q82[6] lines that express N-terminal fragments of HTT, as well asthe YAC128 [7] and BACHD [8] lines that express a mutantversion of the full-length protein. The genetic basis of HD ismore precisely recapitulated by the knock-in models in whichan expanded CAG repeat has been inserted into mouse Httand that develop a more slowly progressing phenotype[9,10,11]. There are several comparable late-stage phenotypesbetween R6/2 mice at 12-14 weeks and HdhQ150 mice at 22

PLOS ONE | www.plosone.org 1 December 2013 | Volume 8 | Issue 12 | e84726

http://creativecommons.org/licenses/by/4.0/

months of age [12,13,14,15,16]. Consequently, we defined theN-terminal mHTT fragments that are present in HdhQ150 braintissue and found that the smallest fragment is exon 1 protein[17]. We have subsequently shown that this is generatedthrough the mis-splicing of exon 1, indicating that the R6 miceare a model for the aberrant splicing that occurs in HD [18].This current study is part of a continuing detailed comparison ofthe progressive pathologies exhibited by the R6 and knock-inmouse models.

We have previously described the concomitant behavioraldysfunction and development of brain abnormalities detectedthrough magnetic resonance imaging (MRI) exhibited by theR6/2 mouse line, expressing ~210 CAG repeats [19]. A similar,but lesser used mouse model, the R6/1 line, expresses fewerCAG repeats (typically ~110-120) and consequently exhibit aslower pathological progression and elongated life-span,compared to R6/2, up to 40 weeks [5]. These featurespotentially make it more suitable and sensitive to investigateearly molecular changes and their relevance to behavioralimpairments. R6/1 mice exhibit age-related progressiveimpairments at both motor and cognitive tests from as early as2 months of age [20], and age related neuronal loss, brainatrophy and mHTT accumulation [21].

Recent efforts have been made to characterize diseaseburden in HD mouse models, including transgenic [20,21],BACHD [22], YAC128 [23,24,25,26], as well as the CAGknock-in mice of varying repeat lengths[27,28,29,30,31,32,33,34,35]. Such descriptions of pathologicalprogression are essential for understanding how theconstruction of an HD mouse can lead to varying phenotypes.Nevertheless, establishing a causal relationship requires alongitudinal assessment of the emergence of behavioraldysfunction with concomitant measurement of brain atrophy byMRI, along with their correlation with cellular and molecularchanges. We here present a correlative analysis of theseconcomitant changes in behavior and pathology in both femaleand male R6/1 mice.

Materials and Methods

Ethics statementAll experimental procedures performed on mice were

approved by the King’s College London Ethical ReviewProcess Committee and carried out under Home Office License70/6445.

AnimalsHemizygous R6/1 mice were bred by backcrossing R6/1

males to (CBA x C57BL/6) F1 females (B6CBAF1/OlaHsd,Harlan Olac, Bicester, UK). Mice were genotyped and the CAGrepeat size was measured, as described previously [15]. TheCAG repeat was 123.19 (± 1.52, SD); the CAG repeat lengthdid not differ between male (123.33 ± 0.866, SD) and female(123.0 ± 2.16) R6/1 mice. Littermates were divided into fourgroups: male wild type (WT, n= 11), male R6/1 (n=9), femaleWT (n=10) and female R6/1 (n=10), an equal ratio betweenmale:female was aimed for. For exact numbers of subjectsincluded at each stage of analysis see Table S1. All

procedures reported here (behavioral, MRI and histological)were conducted on the same cohort of animals.

WT and R6/1 mice were housed under standard animallaboratory conditions. Room temperature was automaticallymaintained at 21°C ± 1°C. Animals were kept on a 12hlight:dark cycle. Mice were group-housed dependent ongender, but genotypes were mixed within the cages. All micehad access to standard cage environmental enrichment(bedding and play tube) and maintained on a standard chowdiet with tap water available ad libitum. Body weight wasmonitored at 6, 10, 14, 18 and 19 weeks of age. Bodytemperature was recorded up to 18 weeks of age through aninfra-red temperature reader (ThermoScan InstantThermometer, Braun) reproducibly positioned under the thorax.An overview of the experimental design is presented in Figure1A.

Behavioral testsMice were exposed to repeated behavioral testing from 6

weeks of age (Figure 1A). Learning in a swimming T-maze wasprobed at both 6 and 15 weeks of age (~2 week procedure atboth ages). Grip strength capacity and behaviors exhibited inan open field were tested twice, at 10 and 19 weeks. Olfactorycapacity and discrimination of social novelty were investigatedat 10 weeks only, and motor coordination and cognition weretested in the rotarod and fear conditioning paradigms at 11 and12 weeks respectively.

Motor-related tasksRotarod. Motor coordination was tested on an accelerating

rotarod at 11 weeks of age, as described previously [19]. Inbrief, mice were individually placed on an accelerating, rotatingbeam (4-40 rpm) for a maximum of 300 sec. Latency to fallfrom the beam (sec) was recorded as a measure of motorperformance. Mice were exposed to three trials per day for fourconsecutive days (day 1 behavior was considered habituationto the test and the data was subsequently discarded). Thewhole apparatus was thoroughly cleaned with 70% industrialmethylated spirit (IMS) between trials.

Open field. General ambulation was probed in an open-fieldarena at 10 and 19 weeks of age. The test was conducted 2 hinto the mouse’s dark cycle, under red-light conditions. Micewere habituated to the test room conditions for 2 h, thenindividually placed into square, plain white arenas (50 x 50 x 50cm, Engineering & Design Plastics Ltd., Cambridge, UK) for 30min, and behavior was recorded through a video camerapositioned above the apparatus. Activity (distance moved, cm)was tracked and later analyzed using EthoVision 7XT software(Noldus, Netherlands). Arenas were thoroughly cleaned with70% IMS between trials.

Grip strength. Grip strength capacity was assessed at both10 and 19 weeks of age. The protocol allowed for theindependent measurement of the strength of the forelimbs only,as well as fore- and hind limbs taken together. Mice were heldby the base of the tail and gently swung to allow them to gripthe wire-mesh grid attached to a grip strength monitor (BiosebIn Vivo Research Instruments). Mice were either allowed to gripwith forelimbs, or fore- and hind limbs together, and then gently

Characterization of R6/1 Mouse Model of HD


pulled away from the apparatus. Maximum tension (g) beforethe mouse released the grid was recorded. For both gripstrength recordings (forelimb only, and fore- and hind limbstogether), the mice were given three attempts and the averageperformance was taken for analysis.

Non-motor-related tasksSwimming T-maze. Cue learning, and the reversal of cue

learning was conducted in a swimming T-maze at 6 and 15weeks of age, using an adapted protocol [24]. The mazeconsisted of a start corridor (122 cm length, 12 cm width, 35cm depth made from black, opaque plastic) leading to a leftand right choice arm (60 cm length, 12 cm width, 35 cm depth)forming a T-junction. The maze was filled with water madeopaque through the addition of a whitening agent (Marvel dried

skimmed milk, Premier Foods, UK). Water was replaced andthe maze thoroughly cleaned every third day. A transparentplastic square escape platform was submerged 0.5 cm,invisible under the water surface. Water temperature wasmaintained at 21-25°C throughout the duration of theprocedure. For “cue learning”, the escape platform waspseudo-randomly located (location order picked randomly out-of-a-hat), at either the left or right end of the top arm and wascued to the presence of an illuminated desktop lamp directlyover the escape platform. Mice were individually placed in thestarting position (at the base of the starting arm, oppositewhere it bisects the top arm) and allowed to swim to the T-junction of the two arms whereupon it made a decision to swimeither toward the light (where it reached the escape platformand was removed from the maze), or away from the light

Figure 1. Experimental design and body weight measures. (A) Experimental design from 6 to 19 weeks of age. All mice weresubjected to repeated assessments of behaviors, alongside in vivo MRI taken at 9 and 17 weeks of age. At 19 weeks of age all micewere culled and an ex vivo MRI scan was taken. STM = swimming T-maze; GS = grip strength; OF = locomotor activity in an openfield, OD = odor discrimination, SI = social interaction, RR = rotarod, FC = cued and contextual fear conditioning. Body weight gainfor both male (B) and female mice (C). Body weight analyzed using three-way ANOVA (Genotype X Gender X Age). Body weight:WT male n=11, R6/1 male n=9, WT female n=10, R6/1 female n=10-8. All data presented as means ± SEM; *p<0.05.doi: 10.1371/journal.pone.0084726.g001



(where it was forced to remain in the incorrect arm for 10 secbefore being allowed to enter the correct side of the top armand find the escape platform). Mice were towel dried betweentrials. Mice were exposed to 12 trails per day until they reachedcriterion (10 correct choices out of 12 successive trials(including rolling over from the previous day), 83.3% correctrate), whereby they began the second part of the trial, termed“cue reversal” learning. For cue reversal training, mice wereexposed to the same conditions to those described for the first,cued learning part of the task, but the desktop lamp cue wasnow situated at the opposite arm to escape platform. Thus,mice were forced to re-learn the escape conditions. Cuereversal training was run until mice reached criterion (again,83.3% correct rate), at which point the assessment wasdiscontinued.

Fear conditioning. Cued and contextual fear conditioningwas used to probe cognitive function in these mice at 12weeks, using an adapted protocol [36]. Experiments wereconducted using a standard fear conditioning system (TSE-Systems, Germany). The test was divided into three days, day1 “training”, day 2 “cue recall and extinction” and day 3 “contextrecall”.

Day 1 “training”. The operant chamber was scented usinga solution of 79.5% water, 19.5% ethanol, 1% vanilla extract.Mice were individually placed in the arenas and exposed to theconditioning protocol; three pairings of a conditioned stimulus(30 sec auditory tone), and an unconditioned stimulus foot-shock (0.5mA, 1 sec duration).

Day 2 “cue recall and extinction”. The operant chamberwas thoroughly cleaned with 50% ethanol to remove all vanillaextract odor. Monochrome patterned wall inserts and a plainwhite floor insert were added to change the context of theoperant chamber. Mice were individually placed in the chamberand subjected to 25 exposures of the conditioned stimulus(auditory tone), but no foot shock, in order to measure cuedrecall and extinction of conditioned fear.

Day 3 “context recall”. The operant chamber was returnedto the identical context as described for Day 1 (chamber insertsremoved, and vanilla odor added) with mice being placed in theoperant chamber for 5 min with no conditioned orunconditioned stimuli to determine contextual recall ofconditioned fear.

Throughout the duration of these tests, behavior wasvideotaped by a camera positioned above the apparatus.Mouse immobility (a complete absence of movement exceptbreathing) was considered a measure of fear, scored by aninvestigator blinded to the experimental groupings. An inter-rater reliability of >95% confidence was achieved. For cuedrecall and extinction on Day 2, immobility over five summed CSexposures (giving a total of 5, 30 sec blocks) was used foranalysis. For contextual recall on Day 3, immobility wasexpressed over the entire 5 min trial.

Odor discrimination. This test was conducted to probeolfactory function of mice at 10 and 19 weeks of age. Micewere habituated to plain white arenas (50 x 50 x 50 cm), for 10min per day for two days prior to starting the test. On the testday, two samples of mouse litter were available forinvestigation over a total of 5 min, one sample being clean

litter, the other being soiled litter from unfamiliar sex-matchedmice. Samples of litter were held in up-turned eppendorf tubeswith 0.5 cm of the tips removed to allow the animal to smell thecontents. Behavior was video taped from above andsubsequently analyzed. Time spent investigating the tubes(nose pokes directly at the aperture of the tubes) wasquantified by an investigator blinded to both the samples andexperimental groupings. Preference for the soiled litter wasexpressed as percentage of time investigating soiled sampleout of total time investigating either sample. An inter-raterreliability of >95% was achieved. Arenas were thoroughlycleaned with 70% IMS between trials.

Social interaction. Social behaviors, as well as the ability todiscriminate novelty, was tested at 10 weeks of age using anadapted protocol [37]. Test mice were habituated to plain whitearenas (50 x 50 x 50 cm) for 1 h. Two corrals (up-turned penholders, Staples, UK) were then positioned inside the arenas inreproducible positions, and the test mice were allowed tohabituate to the corrals for a further 30 min. During “exposure1”, a sex-matched, 8 week old C57Bl/6 mouse was positionedin one corral and the test mouse was allowed to investigate,but not physically interact with the mouse for 5 min. This now“familiar” mouse was removed, and following an inter-trial-interval of 30 min “exposure 2” was conducted, whereby the“familiar” mouse was replaced back into the same corral and adifferent, “novel” C57Bl/6 mouse was placed in the other corral,the test mouse was free to investigate either the familiar ornovel mouse. Behavior was video-taped from above andsubsequently analyzed. Time spent investigating the familiarand/or novel mice was measured by an investigator blinded tothe experimental groups. An inter-rater reliability of >95% wasachieved. Arenas and corrals were thoroughly cleaned with70% IMS between trials.

Magnetic Resonance ImagingIn vivo longitudinal MRI. WT and R6/1 mice were scanned

twice in vivo, at 9 and 17 weeks of age. Mice wereanaesthetized using 5% isoflurane along with a combination ofmedical air (0.7 l/min) and oxygen (0.3 l/min). Once fullyanesthetized, mice were positioned and fixed into a plasticframe, where anesthetic was administered through a facemask.Mice were maintained under anesthesia, typically between1-2% isoflurane for the duration of the scanning. Temperaturewas maintained through a homeostatic heating airflow systemand breathing rate monitored through a respiration balloonpositioned under the thorax (Small Animal Instruments, NewYork, USA). Post scanning, but prior to recovery, mice wereadministered 0.1ml saline i.p. to abate dehydration.

Images were acquired on a 7T horizontal bore MagneticResonance Imaging (MRI) system (Varian, Paolo Alto,California, USA), with a 100 Gauss gradient set insert and a 39mm-bore (transmission and receiver) radiofrequency coil(Rapid, Germany). The scanner was controlled through VnmrJsoftware (Varian, Paolo Alto, California, USA). Correctpositioning of the mouse within the RF coil was confirmedthrough a series of scouting images. A Multi-Echo-Multi-Slice(MEMS) scan was then acquired (TR = 2500 msec, TE = 10msec, echo train = 8, averages = 4, matrix = 128 x 128, FOV =



20 x 20 mm, 30 coronal slices at 0.5 mm thickness, 21 minacquisition time). The typical signal-to-noise ratio (SNR) = 5.4,the typical white:grey matter ratio (WGR) = 1.25. Coronal sliceswere positioned based on a reproducible anatomical marker(the most visibly posterior part of the cerebellum).

Post-acquisition, all eight echoes were summed into a singlestructural image set. As previously described, [19,38], theseimages were used to manually delineate neuroanatomicalstructures in JIM Ver. 5.0 (Xinapse Systems, Alwincle, UK).Regions-of-interest (ROIs) consisted of: whole brain, cortex,striatum, hippocampus, and corpus callosum. ROIs weredelineated by two investigators blinded to the experimentalgroupings, and intra- and inter-rater reliability was consistently≥ 95% confidence. Details of neuroanatomical inclusion criteriaand delineation guidelines were identical to those describedpreviously [19]. All information outside of the ROIs wassubsequently masked out, the ROIs were then individuallysaved in NIFTI format. Volumetric data were calculated andprocessed semi-automatically using Python Ver.2.6 (PythonSoftware Foundation). To measure changes in T2 relaxivity(reflective of tissue composition), maps of T2 signal intensitywere obtained through a mono-exponential fit of the eightechoes. The ROIs were superimposed onto the maps of T2signal intensity allowing for the generation of mean T2relaxation times within each ROI. A small circular ROI wastaken for cheek muscle tissue T2 relaxivity in order to act as aninternal control measure.

Ex vivo MRI. Post-mortem ex vivo MRI scans were taken ata higher-resolution compared to the in vivo scans and did notsuffer from potential artefacts arising in live scanning (e.g. headmovement). A more precise measurement of subtle changes inbrain structures is therefore achievable. Following the finalbehavior test, at 19 weeks of age, mice were anaesthetizedusing terminal anesthesia (Euthatal, Marial, Harlow, UK), andthen transcardially perfused with heparinized saline (50 units/ml), followed by Parafix (4% paraformaldehyde, PioneerResearch Chemical Ltd., Essex, UK). Whole heads wereremoved and submerged in Parafix and kept at 4°C until exvivo imaging. The scanning set-up was identical to that usedfor in vivo imaging. Correct positioning of the mouse headwithin the RF coil was confirmed through a series of scoutingimages. An MEMS sequence was then acquired (TR = 3000msec, TE = 10 msec, echo train = 8, averages = 22, matrix =192 x 192, FOV = 19.2 x 19.2 mm, 35 coronal slices at 0.5 mmthickness, acquisition time for scan was ~3.5 hours). TypicalSNR = 11.96, typical WGR = 1.54. Resulting images weresubsequently converted to ANALYZE 7.5 for Tensor BasedMorphometry (TBM).

Tensor Based Morphometry. An un-biased whole-braincomparison of WT and R6/1 at each imaging time-point wasperformed using an automated image processing pipeline [39].All scans were first registered with 6 degrees of freedom (dof)(to remove differences in position and orientation using a rigid-body assumption) and 9 degrees of freedom (to optionallyremove global differences in scale using a growth model basedon uniform scaling) using a population based approach whichhas proven robust in rodent imaging applications [40]. Theneach scan was non-rigidly registered to the WT mean at the

same time-point using a high-dimensional fluid registrationtechnique [41,42,43]. This technique models the coordinatemapping between scans as the flow of a viscous fluid and cansuccessfully map large structural displacements as well assmaller localized changes in shape. This technique haspreviously been applied in other rodent models, e.g. structuralremodeling in stroke [39]. The fluid registration results in adense displacement field that maps each point in the originalscan to the corresponding point on the reference mean. Fromthis map, an estimate of apparent volume difference (theJacobian determinant) between the scan and the WT mean ateach voxel can be obtained. TBM analysis then applies voxel-wise non-parametric t-tests to these volume differenceestimates to determine the location of statistically significantdifferences in brain tissue volume of R6/1 compared with WT.Significance levels are corrected for multiple comparisonsacross voxels using the False Discovery Rate. The analysisfollowing 6 dof registration finds absolute volume differences(i.e. the fluid registration includes differences in brain size andthe analysis finds absolute volume difference betweenregions), whereas 9 dof registration finds volume differencesrelative to whole brain volume, V (i.e. the fluid registration doesnot include differences in global brain volume and the analysisfinds volume differences relative to whole brain volume).Collectively, these analyses allow for the comparison of WTversus R6/1 at each time point (three image sets in total, 9 and17 weeks in vivo and 19 weeks ex vivo), as well as the age-related change within each genotype (from 9 to 17 weeks ofage). A detailed methodological description of this approachcan be found in [39].

HistologyUpon completion of ex vivo imaging, brains were removed

from the skulls, rinsed in phosphate buffered saline (PBS) andcryoprotected in 30% sucrose in PBS (+0.05% sodium azide)until sectioning. Coronal sections were taken serially at 50 μmthickness on a freezing microtome (HM430 Microm, ThermoScientific) and stored at -20°C in tissue cryoprotective solution(+ 0.05% sodium azide) until staining. Histological processingand data collection was performed identically to that describedpreviously in [19].

Immunohistochemistry. Sections were washed in PBSprior to incubation for 30 min in 3% H2O2 in PBS to quenchendogenous peroxidase activity. Non-specific binding wasblocked with a 1 h incubation in 10% normal serum with 0.3%Triton X-100 in PBS. Sections were then incubated overnight at4°C in primary antibodies against NeuN (1:500, Millipore,Watford, UK) or S830 (1:2000), raised against exon 1 HTT with53 glutamines [44], prior to incubation in appropriatebiotinylated secondary antibody (Vector, Peterborough, UK) for2 h at RT, followed by 1 h incubation in an avidin-biotinylated-peroxide complex (1:100, Vector, Northampton, UK). 3, 3'-diaminobenzidine (Sigma-Aldrich, Poole, UK) was used as thechromagen.

Cortical Thickness. Assessment of regional corticalatrophy was determined by thickness measurements of primarymotor cortex (M1) and primary sensory cortex (S1) on NeuN-stained sections. In each region, 10 vertical lines were drawn



covering all layers from the most dorsal horn of the corpuscallosum to the pial surface. The mean length taken was for 3consecutive sections caudally from when the corpus callosumbridges across the two hemispheres (approximately Bregma+1.10 mm).

Stereology of NeuN-stained sections. Unbiasedstereological estimates of volume and neuronal number wereobtained using StereoInvestigator software (Microbrightfield,Willston, VT). All stereological measurements were performedwith the observer being blinded to the experimental grouping.The Cavalieri method was used to obtain unbiased estimatesof striatal and M1 cortical reference volumes [45]. ROIs weredefined by x1.6 magnification lens through reference toneuroanatomical landmarks. For both the striatum and M1cortex, equally spaced sections (50 μm thickness each,typically 4-6 sections with a 450 μm gap) were analyzed (seeFigure S1 for boundaries drawn on histological sections).

As defined by Sadikot and Sasseville [46], for striatalmeasurements, sections were sampled anteriorly from the firstappearance of the genu of the corpus callosum (Bregma = +1.1mm) to the first evidence of a hippocampal formationposteriorly (Bregma = -0.94 mm). The dorsal and lateralboundaries consisted of the corpus callosum with the medialboundary being the lateral ventricles/external capsule. Forsections rostral to where the dorsal 3rd ventricle has joined thelateral ventricles, ventral boundaries become lateral ventricles/globus pallidus. The striatal volume was sampled on 4-5sections for both WT and R6/1.

M1 cortex was measured anteriorly from +1.1mm Bregma toposteriorly -0.94 mm Bregma, from layers II to VI, as defined inthe stereotaxic atlas [47]. The absence of cortical layer IV(indicative of the S1 cortex) defined the lateral boundaries ofM1, whereas medial boundaries consisted of the most dorsalpart of the corpus callosum. M1 was sampled on 4-5 sectionsfor both WT and R6/1.

To obtain unbiased estimates of neuronal numbers, theoptical fractionator was employed as a stereological probe(coefficient of error <0.1). Section thickness and neuronalcounts were performed under oil immersion with the x100objective (Zeiss) with a numerical aperture of 1.4. A samplinggrid was applied appropriate to the structure measured (cortex= 200 μm x 200 μm, striatum = 400 μm x 400 μm) with acounting frame of 65 μm x 35 μm with a mean thickness of 18μm. Guard zones of 0.5 μm were applied at the top and thebottom of each frame with a mean dissector height of 17 μm.

Quantitative analysis of S830. Evaluation of mHTTimmunoreactivity in different brain regions was performed usingan intensity-based measurement of S830 staining. Non-overlapping images (using fixed exposure and light intensitiesat x40) were obtained from 3 consecutive sections expressingthe striatum or hippocampus, and 6 sections for the cortex. Intotal, 30 striatal (10 per section), 60 cortical (10 per section)and 36 hippocampal subregion (12 per section) images weretaken. All images were captured in RGB using a live videocamera (JVC, 3CCD, KY-F55B), mounted onto a ZeissAxioplan microscope.

Staining intensity was quantified using threshold-basedanalysis software (Image Pro Plus, Media Cybernetics, IL,

USA) assessing optical density of the immunoreactive product.Threshold levels were chosen based on the minimum level oftransmitted light needed to detect the immunoreactive producton a scale of 0 (100% transmitted light) and 255 (0%transmitted light) for each pixel. Two levels were taken tomeasure dense, nuclear mHTT inclusions (90), and totalaggregated mHTT staining (nuclear and extra-nuclear, 130);mean percentage immunoreactivity area per field of view (FOV)was recorded.

Statistical analysesData was graphed using Prism Ver.5.0b (GraphPad

Software, California, USA). Statistical analyses were calculatedusing SPSS Statistics Ver.20 (IBM, Portsmouth, UK). All datawere screened for statistical outliers using Grubbs’ Test(GraphPad Software, California, USA). Grubbs’ test calculatesa z ratio for each value within a given dataset (determinedthrough subtracting each value from the group mean, anddividing it by the standard deviation). If the z ratio for any givenvalue is greater than those proposed by Grubbs for thatpopulation size, it is considered an outlier. If this was the case,the value was excluded from analysis. Due to either animalloss (total of 2 cases, female R6/1 only), or occasional missingdata samples and removal of statistical outliers (total absentdata samples for male WT = 15, male R6/1 = 8, female WT = 8,female R6/1 = 20), the number of animals varied for each testat different time points. A table summarizing the number ofsubjects included for each measure and the reason forexclusion can be found in Table S1.

The acquired datasets can be separated into those that werecollected longitudinally, those acquired at a single time point,correlational analyses, as well as Tensor Based Morphometrystatistics.

Longitudinal datasets. These datasets refer to tests wheredata was collected at more than one time point and include:body weight assessment, body temperature assessment,locomotor activity, grip strength measurements, behaviors in aswimming T-maze and measurements of volumetry and T2relaxivity through MRI. It was not possible to compute arepeated measures ANOVA due to missing values at differentmeasurement times. To probe the influence and interaction ofgenotype and gender across time, a three-way ANOVA wasapplied (Genotype, Gender and Age, as between-subjectfactors). To determine whether mice exhibit cued fearconditioning extinction (i.e. a form of re-learning that theconditioned stimulus is no longer aversive with repeatedexposure), a repeated measures ANOVA was used (repeatedTone (conditioned stimulus) exposure as within-subject factor,Genotype and Gender as between-subject factors). Post-hoctests with a Bonferroni Correction for multiple comparisonswere applied where appropriate. All main effects from ANOVAscan be found in Table S2.

Datasets taken at a single time-point. These data refer totests that were only conducted once during the study andinclude: rotarod, odor discrimination, social interaction, fearconditioning, and histological measures taken throughstereology. To probe the influence of genotype and gender onthese measures, a two-way ANOVA was applied (Genotype



and Gender as between-subject factors). Post-hoc tests with aBonferroni Correction for multiple comparisons were appliedwhere appropriate. All main effects from ANOVAs can be foundin Table S2.

For the histological quantification of mHTT, there was anabsence of immunoreactivity in the WT mice. This marker,within this group, was therefore not included in the analysis. Tocompare levels of mHTT across various brain regions, andprobe the influence of gender on this measure a two-wayANOVA was applied (Region and Gender as between-subjectfactors). A post-hoc test was conducted for multiplecomparisons, with a Bonferroni Correction, comparing maleand female R6/1 mice at each brain region.

Correlative analyses. The Pearson Correlation Coefficientwas used for all correlative analyses. To account for multiplecorrelations of the same data, the significance level wascorrected using a Bonferroni Correction by dividing thestandard p value (0.05) by the number of comparisons made,resulting in an appropriately adjusted p value for each dataset.

Tensor Based Morphometry Statstics. As described in[39], a non-parametric two-tailed t-statistic, assuming unequalvariance between groups, is computed at each voxel(approximately 42,000) in the brain; permutation-testing is usedto assess significance. The effective number of permutations ateach voxel is increased by pooling null-distribution statisticsfrom other voxels to allow for accurate multiple comparisonscorrection using the False Discovery Rate [48]. The minimumnumber of required permutations is approximately (number-of-voxels / FDR-significance-level) = 42,000 / 0.05 = 840,000.

Results

Physiological measuresAbnormal weight gain is a common feature of mouse models

of HD. Both male (Figure 1B) and female (Figure 1C) R6/1mice had lower body weights at later ages compared to WTs[F(Genotype X Age)4,195 =6.173, p<0.001]. Although there wasno interaction between gender and genotype [F(Genotype XGender)1,195=0.845, ns] weight loss reached statisticalsignificance for male mice at 19 weeks only (p=0.025, multiplecomparison with Bonferroni Correction). Overall, differences inbody temperature were detectable between WT and R6/1 mice(data not shown; [F(Genotype)1,157=5.768, p=0.018]), but therewas no interaction with gender [F(Genotype XGender)1,157=0.18, ns] or age [F(Genotype X Age)3,157=0.293,ns]. Moreover, upon post-hoc testing, when corrected formultiple comparisons, there were no genotype-dependentdifferences detected at any given time point.

R6/1 exhibit age-related motor and cognitive deficitsThe presence of motor abnormalities is an important feature

of HD mouse models, here we tested WT and R6/1 mice atthree tasks designed to probe motor ability: rotarod, locomotoractivity and grip strength.

Motor-related-tasks. Rotarod was tested at 11 weeks ofage only. There was an overall difference between the twogenotypes [F(Genotype)1,39=20.794, p<0.001], but nointeraction with gender [F(Genotype X Gender)1,39 =3.918, ns].

Despite having a lower performance compared to WTs at thistask, male R6/1s did not reach statistical significance (Figure2A), whereas, female R6/1s were significantly impaired versusWTs (p<0.001, Figure 2B).

Exploratory activity in the open field was assessed during thedark cycle, at both 10 and 19 weeks of age. Overall, WT andR6/1 mice behaved differently during this task[F(Genotype)1,76=7.296, p=0.009], but there was no interactionwith either gender [F(Genotype X Gender)1,76=2.115, ns] or age[F(Genotype X Age)1,76=0.443, ns]. After correcting for multiplecomparisons, exploratory activity exhibited by either male(Figure 2C) or female (Figure 2D) R6/1 mice was not differentto WT controls at either age studied.

Grip strength capacity (forelimbs; simultaneous fore- andhind limbs) was assessed at both 10 and 19 weeks of age.Forelimb grip strength capacity of R6/1 mice changed over time[F(Genotype X Age)1,77=6.712, p=0.012], but this was notinfluenced by gender [F(Genotype X Gender)1,77=0.1, ns].Maximum tension exhibited by both male (Figure 2E) andfemale R6/1s (Figure 2F) dropped with age, with only malesbeing statistically different at 19 weeks (p=0.015). The gripstrength of the fore- and hind limbs, when taken together,followed a similar pattern to that of the forelimbs only, as therewas an age-related change in performance between WTs andR6/1s [F(Genotype X Age)1,77=13.107, p=0.001], and, again,gender did not exert a significant influence [F(Genotype XGender)1,77=1.033, ns]. Maximum tension of the fore and hindlimbs of the R6/1s dropped to below that of WTs at 19 weeks ofage for both males (p=0.003, Figure 2G) and females(p=0.016, Figure 2H).

Non-motor-related tasks. Clinical signs of HD are notrestricted to motor abnormalities and these features shouldalso be recapitulated in mouse models for greater face validity.To probe such non-motor behaviors, mice were exposed tolearning and memory paradigms (swimming T-maze and fearconditioning), as well as measures of olfactory function (odordiscrimination) and recognition of social novelty (socialinteraction).

The swimming T-maze was used for the assessment of cuedlearning and its subsequent reversal at both 6 and 15 weeks ofage. Overall, the ability to learn a visual cue signifying anescape route (termed “cue learning”) was different between WTand R6/1 mice [F(Genotype)1,77 =5.066, p=0.028]. Despitethere not being a significant interaction with either gender[F(Genotype X Gender)1,77=0.185, ns] or age [F(Genotype XAge)1,77=2.663, ns], male R6/1s were not significantly differentto WT controls at either age (Figure 3A), but female R6/1 micetook longer to learn the rule compared to female WTs at 15weeks only (p=0.01, Figure 3B). Reversing the previouslylearned cue-based rule, termed “cue reversal”, is a morecognitively demanding task and this was reflected by thegenerally longer trials required to reach criterion (on average46.01% longer compared to the more simple cue learning).There was an age-related difference in cue reversal learningbetween WT and R6/1 mice [F(Genotype X Age)1,75=19.17,p<0.001]. Similar to the cue learning phase of this task, genderdid not influence behavior at cue reversal [F(Genotype XGender)1,75=0.348, ns]. Post-hoc analyses revealed that at 6



Figure 2. Performance at motor-related tasks. Both male (A) and female (B) R6/1 mice exhibited lower latency to fall from anaccelerating rotarod at 11 weeks of age, reaching statistical significance in the females only. Conversely there was no difference inlocomotor activity in an open field between WT and R6/1 mice at either 10 or 19 weeks of age (C & D). Forelimb grip strength (E &F) declined with age in R6/1 mice, as did both fore- and hind limb strength when taken together (G & H). Rotarod analyzed usingtwo-way ANOVA (Genotype X Gender); locomotor activity and grip strength analyzed using three-way ANOVA (Genotype X GenderX Age). Rotarod: WT male n=11, R6/1 male n=9, WT female n=10, R6/1 female n=10; Locomotor activity: WT male n=11, R6/1male n=9, WT female n=10, R6/1 female n=8-9; Grip strength: WT male n=11, R6/1 male n=9, WT female n=10, R6/1 femalen=8-10. All data presented as means ± SEM; *p<0.05, **p<0.01, ***p<0.001.doi: 10.1371/journal.pone.0084726.g002



weeks of age R6/1 mice had a similar learning capacity toWTs, but by 15 weeks both male (p=0.018, Figure 3C) andfemale R6/1s (p=0.006, Figure 3D) took longer to learn this rulecompared to WT controls.

Fear conditioning relies upon the formation of a conditionedfear response to an auditory tone and salient context, andprovides a measure of memory retention and extinction.Performance at this task was investigated at 12 weeks of agein WTs and R6/1s. On the training day, immediately followingthe conditioning foot-shock exposure, there was no differencein immobility (fear expression) between the WT and R6/1 mice(data not shown; [F(Genotype)1,35=0.193, ns]) and there was noinfluence of gender [F(Genotype X Gender)1,35=0.436, ns],indicating a similar response to the conditioning protocol, and,hence, memory acquisition. On the second day, all animalswere subjected to repeated exposure of the conditioning tonewithout foot-shock administration to probe memoryperformance. Initially mice expressed a robust fear response,and, with repeated exposure to the tone, mice exhibited lessimmobility ([F(Tone)4,27=6.879, p=0.001]; Figure 3E&F),indicative of cue extinction learning. There was no interactionbetween Genotype and the response to the repeated Tone[F(Genotype X Tone)1,27=1.116, ns] suggesting a similar patternof extinction learning between WTs and R6/1s. However,overall, fear response across all these tones was significantlylower in R6/1s compared to WTs [F(Genotype)1,30=6.824,p=0.014], indicating some abnormality in R6/1 performance atthis task, but there was no influence of gender on this behavior[F(Genotype X Gender)1,30=0.042, ns]. On the third day of thetask, contextual recall was tested over 5 min. Overall R6/1mice expressed less fear response compared to WTs([F(Genotype)1,34=6.006, p=0.02]; Figure 3G&H), but this wasnot influenced by gender [F(Genotype X Gender)1,34=0.004, ns].These results imply an impaired memory performance in theR6/1 mice at 12 weeks, but the exact nature of the deficitremains unclear.

Olfaction was tested in the mice at 10 weeks of age usingthe odor discrimination task (Figure 3I). There was nodifference in WT and R6/1 at the performance of this task[F(Genotype)1,35=0.349, ns] and no effect of gender[F(Gender)1,35 =0.204, ns]. Following the odor discriminationtask, mice were tested at their social interaction and ability todetermine novelty at 10 weeks. On the first part of the task,time spent investigating an unfamiliar mouse was similarbetween WTs and R6/1s [F(Genotype)1,39=0.936, ns], whichwould suggest unaffected social behaviors in the R6/1s. At thesecond part of the task (Figure 3J), (exposure to the “familiar”mouse from the first trial along with a “novel mouse”), therewas no difference in preference for novelty between WT andR6/1 [F(Genotype)1,38=0.253, ns], and no influence of gender[F(Gender)1,38=0.704, ns].

To determine interactions and associations of performanceat these behavioral tasks (both motor- and non-motor-related),a correlative analysis was performed. These analyses wereseparated by age tested; behaviors measured between 6-12weeks (Table S3) and behaviors measured between 15-19weeks (Table S4). There were no significant correlationsbetween the performance at the various behavioral tasks

recorded between 6 and 12 weeks of age. However, at 15 to19 weeks, a higher grip strength in the forelimbs only wasassociated with higher grip strength in both the fore- and hindlimbs for the males (r=0.79, p<0.005). A longer time to learnthe cue reversal task in the swimming T-maze was onlyassociated with lower fore- and hind limb grip strength infemales (r=-0.669, p<0.005). These effects were only evident,however, when considering both WT and R6/1 mice together,and, therefore likely reflect the concomitant development ofimpairments in two measures in R6/1 mice; as performance intwo measures declines over time, a strong correlation is formed(Figure 4A&B).

R6/1 exhibit progressive brain atrophyAlongside the detailed analysis of behavioral dysfunction in

R6/1 mice, longitudinal in vivo MRI was conducted at 9 and 17weeks of age to investigate changes in neuroanatomy (Figure5A). The striatum, hippocampus, cortex, corpus callosum andwhole brains were manually delineated and regional volumetryand T2 relaxivity were measured in the same ROIs.

Longitudinal MR volumetry provides a measure of volumechange in the R6/1 mouse brain (Figure 5B). The whole brainvolume of R6/1 mice was different to WTs over time[F(Genotype X Age)1,78=7.501, p=0.008]. In fact, R6/1s werealready 94.5% that of the WTs at 9 weeks of age, reachingstatistical significance for both the males (p=0.023) andfemales (p=0.044), although there was no overall interactionbetween genotype and gender [F(Genotype XGender)1,78=0.138, ns]. This difference was amplified at 17weeks to 89.79%. Overall striatal [F(Genotype)1,78=26.481,p<0.001], cortical [F(Genotype)1,78=61.667, p<0.001] andhippocampal volumes [F(Genotype)1,78=64.687, p<0.001] weredifferent between WT and R6/1 mice, but this was notinfluenced by gender (striatum [F(Genotype XGender)1,78=0.464, ns]; cortex [F(Genotype XGender)1,78=0.033, ns]; hippocampus [F(Genotype XGender)1,78=0.056, ns]). There was a significant interactionbetween genotype and age in the cortex [F(Genotype XAge)1,78=5.224, p=0.025] and hippocampus [F(Genotype XAge)1,78=11.447, p=0.001], but, interestingly, this interactionwas not evident for the striatum [F(Genotype X Age)1,78=3.019,ns]. A Bonferroni-corrected multiple comparison analysisrevealed no differences in these regional volumes at 9 weeksof age, but by 17 weeks of age all three regions weresignificantly smaller in the R6/1s: the R6/1 striatum was82.36% that of WT (males p=0.008, females p=0.004); theR6/1 cortex was 84.91% that of WT (males p<0.001, femalesp<0.001); the R6/1 hippocampus was 82.49% that of WT(males p<0.001, females p<0.001). There was no detectabledifference in corpus callosum volume at either 9 or 17 weeks ofage. Whole brain volume was significantly correlated withcortical (r=0.708, p<0.0011) and hippocampal (r=0.693,p<0.0011) volumes in female R6/1s (Table S5), suggesting arelationship between changes in whole brain and sub-regionalvolumes.

As well as measures of brain volumes, T2 relaxivitymeasures (reflective of tissue composition) were taken in thestriatum, cortex, hippocampus, corpus callosum and muscle



Figure 3. Performance at non-motor-related tasks. Male R6/1 mice did not exhibit any impairment in cue learning in theswimming T-maze (A), whereas female R6/1 mice were deficient at 15 weeks of age (B). Both male (C) and female (D) R6/1sdeveloped an age related deficit in cue reversal learning in a swimming T-maze. There was no obvious difference in cued extinctionbehavior between WTs and R6/1s (E & F), or a difference in contextual recall (G & H). At the odor discrimination task (I) and novelsocial interaction test (J) there was no difference between the four groups. Swimming T-maze analyzed using three-way ANOVA(Age X Genotype X Gender); fear conditioning cue extinction analyzed using a repeated measures ANOVA (Genotype X Gender XTone); fear conditioning context recall, odor discrimination and novel social interaction analyzed using a two-way ANOVA (GenotypeX Gender). Swimming T-maze: WT male n=10-11, R6/1 male n=8-9, WT female n=10, R6/1 female n=9-10; Fear conditioning: WTmale n=11, R6/1 male n=9, WT female n=7-8, R6/1 female n=7-8; Odor discrimination: WT male n=11, R6/1 male n=8, WT femalen=9, R6/1 female n=8; Novel social interaction: WT male n=10, R6/1 male n=9, WT female n=10, R6/1 female n=10. All datapresented as means ± SEM; *p<0.05, **p<0.01.doi: 10.1371/journal.pone.0084726.g003



tissue. There were few differences in regional T2 valuesbetween WT and R6/1 mice (data not shown). Overall, T2relaxation times were different between WTs and R6/1s in thestriatum [F(Genotype)1,77=13.294, p=0.001] and corpuscallosum [F(Genotype)1,76=4.403, p=0.04] only, there were nointeractions between genotype with age or gender in anyregion studied. A post-hoc analysis revealed only a shorteningof T2 values between male WT and R6/1 mice at 9 weeks inthe striatum (p=0.002), cortex (p=0.009) and hippocampus(p=0.039). There were no significant effects in the corpuscallosum or muscle tissue.

Compared to the volumetric data, T2 values across thevarious brain regions studied were highly correlated with eachother for both WTs and R6/1s, irrespective of gender (TableS5). This, however, was not the case for the corpus callosumor muscle tissue. These data indicate that changes in T2values are highly consistent and ubiquitous across grey matterirrespective of mouse genotype. There were no significantcorrelations between measures of volumetry and T2 relaxivitywith the exception of a smaller whole brain volume beingassociated with higher T2 relaxation times in the male R6/1sonly (r=-0.756, p<0.0011).

The neuroanatomical basis of behavioral impairments isreflected in associations between these two outcomemeasures, thus a correlative approach to probe associationsbetween brain abnormalities detected through MRI and age-matched functional impairments in behavioral tasks was takenbetween early (between 6 and12 weeks) and later measures(15 and 19 weeks) (Table S6). Only two significantassociations were detected after corrections for multiplecomparisons, both in early measures of behavior and MRI (6 to

12 weeks). When considering both WTs and R6/1s together,longer hippocampal T2 relaxation was associated withhyperactivity in an open field (r=0.702, p<0.001), evident in themales only. Moreover, longer hippocampal T2 was alsoassociated with a lower immobility (memory recall) in the cuedfear conditioning task (r=-0.761, p<0.001), but in the femalesonly.

TBM analysis revealed sub-region specific and gender-dependant differences in the R6/1 brain atrophy

Although informative, manually segmented ROI analysis hassome limitations, such as a lack of sub-regional specificity andthe selective choice of regions to be studied. Tensor BasedMorphometry (TBM), an automated image analysis, allows forthe unbiased detection of changes in individual voxels acrossthe whole brain irrespective of their ROI assignment. TBM wasused to compare WT and R6/1 brains at three time pointsscanned, 9 weeks and 17 weeks in vivo, and 19 weeks ex vivofor males (Figure 6) and females (Figure 7). Although therewas evidence of global brain volume shrinkage (6 dof) in maleR6/1 mice at 9 weeks (Figure 6A), but not female (Figure 7A),there was no specific regional differences when corrected forwhole brain volume change for either gender (9 dof). By 17weeks there was a substantial shrinkage of both whole brainvolume (Figure 6B), as well as sub-regions when corrected forglobal volume change in male mice only. Specifically, theseanalyses revealed subtle expansion of posterior ventricularspaces and, interestingly, in R6/1s, a spared hippocampalvolume. Similar to that detected through the manuallysegmented analysis, the TBM analysis also indicates that themale R6/1 cortex suffered atrophy, but this advanced technique

Figure 4. Concomitant deficits in behavioral tasks result in a robust correlation of those two measures. (A) At 18 weeks ofage R6/1 mice have both lower forelimb, and fore- and hind limb grip strength capacity to a similar degree, separating the data fromthose of WT resulting in a significant, positive correlation (r=0.79, p<0.005). (B) Between 15-19 weeks of age R6/1 mice have lowerfore- and hind limb grip strength, as well as impaired cue reversal learning in a swimming T-maze to a similar degree, forming astrong overall negative correlation (r=-0.669, p<0.005).doi: 10.1371/journal.pone.0084726.g004



allows for a more precise description of this change, as it isapparent that TBM has revealed a global cortical atrophy, butwith a relative sparing of the cingulate cortex (at Bregma~0.26-0.46 mm). These TBM analyses provide more precise

evidence of sub-regional changes that occur within large brainregions, such as the cortex.

Female R6/1s had robust global volume differences (6 dof),but no sub-regional changes (9 dof) were evident even at 17

Figure 5. Longitudinal in vivo MRI. (A) Summed T2-weighted structural group-images from WT and R6/1 mice at 9 and 17weeks of age. (B) R6/1 mice progressive lost regional brain volumes compared to WT controls, with the exception of the corpuscallosum (C. Callosum). Analyzed using a three-way ANOVA (Genotype X Gender X Age). Striatum, cortex, hippocampus, corpuscallosum and whole brain: WT male n=11, R6/1 male n=9, WT female n=10, R6/1 female n=9-10. All data presented as means ±SEM; *p<0.05, **p<0.01, ***p<0.001.doi: 10.1371/journal.pone.0084726.g005



weeks (Figure 7B). These effects are similar, but more robustand better defined, on the analysis of the ex vivo images takenat 19 weeks, where differences between male (Figure 6C) andfemale mice (Figure 7C) were less evident. Interestingly, forboth genders, the greatest magnitude of shrinkage was evidentin the posterior striatum at Bregma ~ -1.22 mm, and substantialcortical atrophy was evident, specifically in the retrosplenialareas, visible at Bregma ~ -2.46 mm.

In order to determine specific age-related alterations withineach separate genotype, a longitudinal analysis was conductedto compare these across time for both male (Figure 8) andfemale (Figure 9) mice. Unexpectedly, male WT mice had noglobal brain volume changes over time, but did show someevidence of specific sub-regional brain changes, principally ashrinkage of the anterior striatum and a ubiquitous corticalatrophy, whist posterior ventricular spaces were enlarged,(Figure 8A), whereas there were no age-related differences infemale WT mice (Figure 9A). Both male (Figure 8B) and female(Figure 9B) R6/1 mice exhibited global brain volume changeand specific regional changes when corrected for wholevolume change. Indeed, both male and female R6/1 miceexhibited a ubiquitous shrinkage of the striatum over time,whereas, within the cortex, the somatosensory cortices weredramatically atrophied at Bregma ~1.22 mm. Interestingly, forboth WT and R6/1 mice, the corpus callosum remained

relatively unchanged with age, consistent with manuallysegmented ROIs.

In summary, both methods of measuring volume changefrom MRI (i.e. manually segmented volumetry and automatedTBM) detected robust shrinkage of grey matter tissue, and arelative sparing of white matter tissue, in R6/1 mice with age.Nevertheless, TBM allowed for a more precise description ofthe anatomical loci of these changes.

R6/1 mice express HD-like histopathological features,including neuronal loss

In order to determine the molecular and neurological basisfor the behavioral dysfunction and brain atrophy observedthrough MRI, a histopathological investigation was conducted.The deposition of mHTT is a neuropathological hallmark of HD.We used the S830 antibody to independently visualize nuclearmHTT inclusions, as well as other forms of aggregated mHTTin both the nucleus and neuropil (referred to here as totalaggregated mHTT) in the R6/1 mouse brain (Figure 10A). S830reactivity was negligible in WT (percentage striatal FOVimmunoreactive was 0.16%), indicating that it reflects aneuropathological marker in the R6/1 genotype. In R6/1s, thelevels of mHTT varied across brain regions for both nuclearmHTT inclusions [F(Region)5,102 =17.808, p<0.001] and totalmHTT aggregates [F(Region)5,104=86.98, p<0.001]. However,

Figure 6. Brain volume changes in male WT and R6/1 assessed through TBM. Maps of local volumetric changes in male R6/1mice compared to male WT 9 weeks in vivo (A), 17 weeks in vivo (B), and at 19 weeks ex vivo (C). Images presented as either bothglobal volume change (6 dof), or region-specific changes corrected for global change (9 dof). Color scales are for volume difference(Volume diff., blue-to-yellow), and the raw t statistical value at each voxel (dark-to-light green). Only volume changes and effects-sizes which survive multiple comparisons corrections across all voxels in the brain (False Discovery Rate with q<0.05) are shown.doi: 10.1371/journal.pone.0084726.g006



there was no influence of gender on either measure (nuclearmHTT: [F(Gender X Region)5,102=0.675, ns]; total mHTT:[F(Gender X Region)5,104=0.503, ns]). Indeed, a post-hocanalysis revealed that there was no gender-dependentdifference between either the nuclear inclusions (Figure 10B)or total aggregated mHTT (Figure 10C) levels in the striatum,cortex or hippocampal subfields (dentate gyrus, CA1, CA2 orCA3). Generally, mHTT levels were positively correlated acrossall brain regions (Table S7). When considering male andfemale R6/1s together, total levels of mHTT were highlycorrelated across the CA1, CA2 and CA3 hippocampalsubfields (r>0.753, p<0.0008), indicating a more ubiquitousdetection in these areas. Interestingly, on TBM analysis of MRI,the hippocampus exhibited little atrophy despite a heavymolecular pathology.

A stereological analysis of sections, stained for the neuronalmarker NeuN (Figure 11A), revealed lower neuronal numbersin R6/1 compared to WTs, evident in both the striatum([F(Genotype)1,37=14.546, p=0.001]; Figure 11B) and M1 cortex([F(Genotype)1,36=15.432, p<0.001]; Figure 11C). There was noinfluence of gender on this drop in R6/1 neuronal number ineither region (striatum: [F(Genotype x Gender)1,37=0.535, ns];M1 cortex: [F(Genotype X Gender)1,36=0.55, ns]). Thesubstantially smaller regional volumes in the R6/1s (striatum by

76.35% of WTs, M1 cortex by 81.35%) accompanied by thisdrop in neuronal numbers (striatum by 80.49% of WTs, M1cortex by 84.29%) resulted in an absence of change inneuronal density in either region (Figure 11D&E; striatum:[F(Genotype)1,37=1.205, ns]; M1 cortex: [F(Genotype)1,36=0.837,ns]). Changes in neuronal number and volume wereaccompanied by substantial cortical thinning (Figure 11F) inboth the M1 cortex [F(Genotype)1,35=92.105, p<0.001] andprimary somatosensory (S1) cortex [F(Genotype)1,35=68.558,p<0.001], but gender had no influence on these measures (M1cortex: [F(Genotype X Gender)1,35=0.55, ns]; S1 cortex[F(Genotype X Gender)1,35=1.347, ns]).

Generally, there was a tendency towards strong positivecorrelations between neuronal number and neuronal density inthe striatum of both WTs (males r=0.949, p<0.0033; femalesr=0.876, p<0.0033) and R6/1s (males r=0.927; p<0.0033,females r=0.825, ns), but not in the M1 cortex (Table S8).Interestingly there was no significant correlation betweenneuronal numbers in the striatum and M1 cortex for R6/1s(males r=0.079, ns; females r=0.638, ns) indicating thatchanges in neuronal number of R6/1 mice may not occurubiquitously across the brain.

Figure 7. Brain volume changes in female WT and R6/1 assessed through TBM. Maps of local volumetric changes in femaleR6/1 mice compared to female WT 9 weeks in vivo (A), 17 weeks in vivo (B), and at 19 weeks ex vivo (C). Images presented aseither both global volume change (6 dof), or region-specific changes corrected for global change (9 dof). Color scales are for volumedifference (Volume diff., blue-to-yellow), and the raw t statistical value at each voxel (dark-to-light green). Only volume changes andeffects-sizes which survive multiple comparisons corrections across all voxels in the brain (False Discovery Rate with q<0.05) areshown.doi: 10.1371/journal.pone.0084726.g007



The histopathological basis of brain abnormalities andbehavioral impairments

By 17 weeks of age, R6/1 mice exhibited substantial brainatrophy visualized through MRI. By combining this imagingalong with measures of histopathology, it is possible toinvestigate potential mechanisms behind this apparent brainvolume loss. Nevertheless, there were no detectableassociations between mHTT load and MRI-based regionalbrain changes (Table S9). Similarly, when considering therelationship between NeuN-based stereological data and MRImeasures of brain abnormalities (Table S10), there were nocorrelations for R6/1 mice. Importantly, there was no

correlation between neuronal numbers in either the striatum orM1 cortex with any MR measures of pathology, which maysuggest that other, non-neuronal histopathological changes inthe brain could have a greater or more direct influence onatrophy. Nevertheless, when considering male WT and R6/1mice together, striatal volume determined through stereologystrongly correlated with most MR-volumetry (striatum r=0.727,p<0.0033; cortex r=0.752, p<0.0033; hippocampus r=0.794,p<0.0033), indicating that indeed these two outcome measureswere probing the same anatomical changes.

Through combining assessments of behavioral performancewith histopathological abnormalities, potential neurobiological

Figure 8. Age-related evolution of brain atrophy in male mice measured through TBM. Maps of local volumetric change inboth male WT (A) and male R6/1 mice (B) between the two time points (9 and 17 weeks of age). Images presented as either globalvolume change (6 dof) or regional-specific changes corrected for global change (9 dof). Color scales are for volume difference(Volume diff., blue-to-yellow) and the raw t statistical value at each voxel (dark-to-light green). Only volume changes and effects-sizes which survive multiple comparisons corrections across all voxels in the brain (False Discovery Rate with q<0.05) are shown.doi: 10.1371/journal.pone.0084726.g008



mechanisms underlying functional deficits in R6/1 mice can beexamined. Levels of mHTT in the brain, however, did notcorrelate with age-matched behavioral performance (TableS11), with the exception of a higher mHTT inclusion level in thehippocampal CA2 being associated with greater impairments inthe swimming T-maze “cue-reversal” learning (r=0.925,p<0.0017). Similarly, behavioral performance did not show anyassociation with stereological measures of pathology in theR6/1s (Table S12), with the exception of a smaller striatum

being associated with an impaired learning at the cue learningphase of the swimming T-maze (r=-0.652, p<0.0033). Takentogether, despite a few correlations of histological measureswith single behavioral findings, there was an absence of directcorrelative associations between these two outcome measures.It is therefore possible that a more complex relationship existsbetween the HD-like neuropathology and subsequentbehavioral impairments in these mice.

Figure 9. Age-related evolution of brain atrophy in female mice measured through TBM. Maps of local volumetric change inboth female WT (A) and female R6/1 mice (B) between the two time points (9 and 17 weeks of age). Images presented as eitherglobal volume change (6 dof) or regional-specific changes corrected for global change (9 dof). Color scales are for volumedifference (Volume diff., blue-to-yellow) and the raw t statistical value at each voxel (dark-to-light green). Only volume changes andeffects-sizes which survive multiple comparisons corrections across all voxels in the brain (False Discovery Rate with q<0.05) areshown.doi: 10.1371/journal.pone.0084726.g009



Figure 10. mHTT accumulation in R6/1 mice. (A) Representative coronal sections from 19 weeks’ old WT and R6/1 mousebrains stained with the S830 antibody for the detection of mHTT. Levels were quantified within seven brain regions using anintensity threshold-based image analysis tool optimized for the detection of nuclear inclusions, or total levels of aggregated mHTT(neuropil aggregates and diffuse nuclear accumulation, highlighted in blue); scale bar 50μm. Field of views (FOVs) were sampled inthe cortex, striatum and hippocampal subfields. Percentage of FOVs positive for S830 stain of nuclear inclusions (B), andpercentage of FOV positive for total aggregated mHTT (C), there were no significant differences between male and female R6/2 foreither assessments. Regional differences in mHTT reflect cellular density. STR = striatum, CTX = cortex, DG = dentate gyrus, CA1= hippocampal CA1 subfield, CA2 = hippocampal CA2 subfield, CA3 = hippocampal CA3 subfield. Analyzed using a two-wayANOVA (Gender X Region). All brain regions investigated: R6/1 male n=8-9, R6/1 female n=7-9. Data presented as means ± SEM.doi: 10.1371/journal.pone.0084726.g010



Figure 11. Histological analyses of NeuN-stained sections. (A) Sample images of 19 week old WT and R6/1 coronal sectionsstained for NeuN; scale bar 50 μm. There was a lower number of neurons in the striatum (B) and M1 cortex (C) in R6/1 mice, but nodifference in neuronal density in these two regions (D & E). There was substantial cortical thinning (F) in both the M1 and S1 corticalsubfields in R6/1 mice. All data analyzed using a two-way ANOVA (Genotype X Gender). Striatal measures of stereology: WT malen=10, R6/1 male n=9, WT female n=10, R6/1 female n=9; M1 cortical measures of stereology: WT male n=10, R6/1 male n=8, WTfemale n=10, R6/1 female n=9; cortical thickness in M1: WT male n=9, R6/1 male n=8, WT female n=10, R6/1 female n=9, and S1:WT male n=9, R6/1 male n=9, WT female n=9, R6/1 female n=9. All data presented as means ± SEM; *p<0.05, ***p<0.001.doi: 10.1371/journal.pone.0084726.g011



Discussion

Establishing how molecular pathology in Huntington’sdisease (HD) leads to structural brain changes and theemergence of functional abnormalities remains a majorchallenge, both clinically in HD sufferers, and preclinically inmouse models of the disease. Moreover, for HD animals to beeffective models in discovering novel therapeutic strategies, weneed an understanding of how the disease-like phenotypesdevelop and relate to one another.

Using an interdisciplinary approach, we have provided adetailed description of the progressive behavioral phenotype,brain atrophy and neuropathological abnormalities, akin tothose observed in HD, in the transgenic R6/1 mouse model.The key findings were 1) R6/1 mice exhibited age-relateddeficits in both motor and cognitive behaviors detectable by 11weeks of age; 2) MRI revealed the first changes in the R6/1brain preceding behavioral dysfunction, as there was a smallerwhole brain volume already at 9 weeks of age, althoughatrophy in various brain regions was only apparent later at 15weeks; 3) accompanying these behavioral and brainabnormalities were typical HD histopathological features,including mHTT accumulation and neuronal loss; 4) there was,however, little evidence that there is a strong associationbetween these typical histopathological markers of HD andbehavioral deficits suggesting that other neurobiologicalchanges might play an important role in the development of theimpairments observed in HD.

R6/1 mice mimic key clinical HD signsGiven the multi-faceted nature of clinical HD phenotypes, it is

essential to thoroughly characterize the different animal modelsaiming to replicate the progressive pathology, as well as thebehavioral phenotype of HD. To characterize behavioralchanges in the R6/1 mouse model of HD, a battery of testsaimed at probing motor and cognitive capabilities was appliedto investigate the emergence of deficits across differentfunctions.

The earliest detectable abnormality was a motor coordinationimpairment on the rotarod at 11 weeks of age, although it isimportant to note that this was the earliest age tested and it istherefore possible that this deficit even precedes this age. Thisearly impairment is consistent with a deficit detected in aprevious study even earlier, at 2 months of age [20]. This motorcoordination task is therefore a useful marker for early HD-likephenotype detection, detectable even before overt body weightloss. Interestingly, this task was found to be impaired in theR6/2 at a similar age, developing around 9 and 13 weeks [19].We also noted robust decreases in grip strength, whichdeveloped between 10 and 19 weeks of age. To ourknowledge, this is the first description of the progressivedevelopment of grip strength deficits in this mouse line,although it has been well described in R6/2 mice [49,50,51].These impairments reflect the decline of neuromuscularfunction and muscle strength, common features of HD. Generalactivity in an open-field, however, was unaltered in R6/1 mice.These data are consistent with those showing an absence ofaltered R6/1 activity measures over 30 min up to 14 week of

age [52]. Yet, others reported diminished exploratory activity asa late (23 weeks, [53]), as well as an early phenomenon (6weeks, [54]) in R6/1 mice. Longer (24h) assessments oflocomotor activity revealed also evidence of hypoactivity from18 weeks of age [55]. Comparing activity assessments in R6/1sfrom various literatures is nevertheless difficult given the grossinconsistencies of testing conditions.

In addition to motor impairments, R6/1 mice developeddeficits in visuo-spatial cue-based learning, and its subsequentreversal, at 15 weeks of age. The neuroanatomicalmechanisms underlying these two distinct forms of learning arecomplex, but the reversal of cued learning requires a higherdegree of cognitive flexibility. Consequently, this is morecognitively demanding, as reflected by the length of time takento acquire the new rule. Orbital and prefrontal cortex [56], aswell as frontostriatal brain regions [57] have been implicated invariants of reversal learning in mice. Abnormalities in theneuronal functioning within these regions have been stipulatedto be involved in reversal learning deficits in several HD mousemodels, including the R6/2 line [58,59,60]. Consistent withthese observations, we found that smaller striatal volumeswere associated with cue-reversal learning deficits in femaleR6/1s.

Neuropathological correlates of behaviorAs expected, we detected the extensive accumulation of

mHTT in the R6/1 mouse brain. As in the R6/2 mouse line [19],there were no differences in protein abundance between maleand female R6/1 mice. It is worthy of note that the relativeamount of mHTT in the R6/1 brain regions investigated iscomparable to those in the R6/2. In the current study, wequantified mHTT levels at 19 weeks of age only, however, aprevious characterization has demonstrated mHTT presence inthe R6/1 brain from as early as 2 months, which greatlyincreased by 4 months [21].

The R6/1 and R6/2 lines had other similarities in brainpathology expression, as both models exhibited a reduction inbrain volumes over time, detected here using longitudinal invivo MRI. As with the R6/2 mice [19], the R6/1 striatumappears not to shrink over time, but rather does not grow at thesame rate as WTs. A similar effect of this absence of normalstriatal growth in the R6/1, rather than regional shrinkage, hasbeen detected previously through histological methods [21].Interestingly, unlike the striatal volume that remains relativelystable from 9 to 17 weeks (only a 4.19 % drop), the corticalvolumes appear to shrink more significantly over time (13.24%drop). Again, these patterns of regional brain volume changeare consistent with those described in the R6/2 mice previously[19]. These regional patterns of brain changes are distinct fromthose generally observed in clinical investigations oflongitudinal brain atrophy in HD. MRI has detected progressiveatrophy of both the basal ganglia and neocortex [61], but othershave shown progressive atrophy of the striatum in patients with>10 years of HD onset, and this progresses with disease[62,63], although general grey-matter changes could not bedetected as early [62]. Indeed, striatal volume loss was a goodpredictor of disease diagnosis in a separate study [64].Therefore, unlike our recent data from mouse models, the



evidence from clinical studies of HD indicate a robust, earlychange in striatal volume in HD sufferers which continues todevelop as the disease progresses.

Alongside manually segmented ROI volumetry studies, TBMallowed an unbiased, sensitive, assessment of volume changeacross the whole brain, including sub-regional changes [65,66].Interestingly, TBM revealed a more substantial pathology in themale R6/1 mice compared to females, a phenomenon that wasless easy to resolve through manual segmentation.Collectively, both manually defined MR volumetry and TBMdemonstrated age-related loss of grey matter, whilst sparingwhite matter in the R6/1 mice. However, TBM allowed for moreprecise detection of shrinkage of anatomical loci within largerbrain regions, such as retrosplenial areas alongside a relativelyspared anterior cingulate within the cortex. Interestingly,although there was a general widespread atrophy occurringacross the whole brain in R6/1, the hippocampus was generallyspared, despite this region being heavily impacted by mHTT.

Despite various similarities with R6/2 mice, there were alsoseveral key differences in the pathological progression of R6/1mice. One major difference between the brain pathology of theR6/1 and R6/2 mice is the presence of neuronal loss in R6/1s.Previously, we have detected no difference in neuronal numberin R6/2 mice at 14 weeks of age [19], however, we detectedneuronal loss in the brain of R6/1 mice at 19 weeks of age,consistent with a previous study which also described neuronalloss in R6/1 [21]. There were no correlations between neuronalloss and brain region volume change in the R6/1 mice, whichmay indicate that non-neuronal changes are responsible for thegross brain volume change in the R6/1. Such changes mayinclude loss of microglia and extracellular matrix molecule, bothbeing features of R6/2 mice [67,68]. Neuronal density washighly elevated in R6/2 mice [19], which was likely a result of alack of neuronal loss coupled with restricted striatal growth.Conversely, in the current study R6/1 mice exhibit bothneuronal loss and volume differences compared to WT mice,thus resulting in a normal cellular density. Heightened cellulardensity in the R6/2 mice may have resulted in lowered tissuewater content, which would, as a result, shorten T2 relaxationvalues. It is therefore conceivable that the absence of cellulardensity change in the R6/1s, results in normal tissue watercontent levels and hence, the generally unaffected tissue T2relaxation values described here.

To determine whether there were clear neuropathologicalsubstrates responsible for the robust behavioral dysfunction, acorrelative analysis was performed between these twooutcome measures. There were few associations betweenbrain abnormalities determined through MRI and age-matchedbehavioral performance, although both progressively worsenedwith age in the R6/1. Studies have revealed that abnormalneuronal functioning underlies behavioral dysfunction in theR6/2 mouse line [69,70], and a disrupted neuronal activity hasalso been reported as in R6/1 mice [71]. It is therefore possiblethat abnormalities in neuronal activity, rather than brain tissueloss, underlie the functional phenotypes in R6 mice.

Intensive behavioral testing may be detrimental tolongitudinal studies

R6/1 mice typically survive until around 40 weeks of age [5],and can be behaviorally characterized up to 7 months of age[20]. The current investigation took the R6/1 mice to 18 weeksof age, whereby they reached a predetermined humaneendpoint approximately 12 weeks earlier than anticipated. Theexperiment was consequently stopped. Given the nature of theintense behavioral characterization undergone here, as well asthe repeated anesthesia for MRI, this may be in partresponsible for an anticipation of the humane endpoint. It ishence important to consider the here presented results in thiscontext.

As detailed in the experimental design, mice were subjectedfirst to repeated swimming exposure followed by an intensefour weeks of behavioral testing. Alongside our test animals, aseparate cohort of WT and R6/1 mice were aged for tissuecollection. Body weight, a common, albeit superficial measureof animal welfare, was lower in the animals exposed to theintense behavioral exposure for both WTs (10.85% lower) andR6/1s (13.58% lower). The increase in general activity andexercise during behavioral testing is likely to reduce bodyweight at the same caloric intake.

Unexpected and uncontrollable stressors have historicallybeen used to model many functional deficits associated withdepression in laboratory animals [72] and standard paradigmsaimed at probing behavior in animals can themselves inducestates of anxiety and emotional stress [73]. Indeed, elevatedlevels of the stress-response hormone, corticosterone,interferes with normal cognitive functioning [74]. This is ofparticular concern when studying animal models of braindisorders given the well described interactions betweenemotional stressors and neurodegenerative disease [75]. Infact, it is possible that transgenic models of HD are moresusceptible to stress-related insults given a hyperactivity of thestress-response pathway, the hypothalamic-pituitary-adrenal(HPA) axis, as has been reported in both R6/2 [76] and R6/1[77] mice. The extent, intensity and duration of behavioralcharacterization of R6 mice (regardless of any possibleheightened susceptibility to stress-related insults) shouldtherefore be carefully considered, as there is a risk of modifyingdisease progression.

Conclusions

This study demonstrated both motor and cognitive deficits inthe R6/1 mouse line with concomitant brain atrophy asdetermined through MRI. Accumulation of mHTT and a loss ofneurons in key brain structures was also evident post-mortem,indicating that neurobiological changes underlie changes inbrain structure and behavior. However, there was littleevidence of a direct link between brain pathology andbehavioral deficits. The R6/1 mouse line therefore replicatesseveral key behavioral and neuropathological features of HD,but presently no mechanistic link between these could beestablished. Investigation of the early stages of behavioralimpairments in this model combined with a more extensivehistopathological and molecular analysis might be required to



establish specific associations that link molecular pathology toa specific behavioral deficit.

Supporting Information

Figure S1. Striatal and M1 cortical inclusion criteria forstereology. Sample coronal brain sections stained with NeuNwith guides of neuroanatomical boundaries for inclusion of thestriatum and M1 cortex for stereological analysis.(TIF)

Table S1. Number of animals used. All tests were conductedon the same cohort of animals. However, there was a variablesubject number for each test due to either death during thestudy, to the occasional missing data sample or to theexclusion of statistical outliers. RR = rotarod, LMA = locomotoractivity in an open field, GS FL = grip strength of the fore-limbs,GS 4L = grip strength of the fore- and hind limbs, TM CL =swimming T-maze cue learning, TM CR = swimming T-mazecue reversal, FC CS = fear conditioning cue recall (totalimmobility over 25 cue exposures), FC CT = fear conditioningcontextual recall, OD = odor discrimination, SI = socialinteraction, STR = striatum, CTX = cortex, HIPP =hippocampus, CC = corpus callosum, WB = whole brain,MUSC = muscle tissue, DG = dentate gyrus, CA1 =hippocampal CA1 subfield, CA2 = hippocampal CA2 subfield,CA3 = hippocampal CA3 subfield, M1 CTX = M1 cortex, S1CTX = S1 cortex.(PDF)

Table S2. Main effects derived from statistical analyses.Main effects derived from two-way and three-way ANOVAs.LMA = locomotor activity in an open field, GS FL = gripstrength of the forelimbs, GS 4L = grip strength of the fore- andhind limbs, TM CL = swimming T-maze cue learning, TM CR =swimming T-maze cue reversal, RR = rotarod, OD = odordiscrimination, SI = social interaction, FC CS = fearconditioning cue recall (total immobility over 25 cueexposures), FC CT = fear conditioning contextual recall, STR =striatum, CTX = cortex, HIPP = hippocampus, CC = corpuscallosum, WB = whole brain, MUSC = muscle tissue, DG =dentate gyrus, CA1 = hippocampal CA1 subfield, CA2 =hippocampal CA2 subfield, CA3 = hippocampal CA3 subfield,M1 CTX = M1 cortex, S1 CTX = S1 cortex.(PDF)

Table S3. Correlations of behavioral measures takenbetween 6 and 12 weeks. Correlations of performance atbehavioral tasks tested between 6 and 12 weeks, presented asPearson r values. RR = rotarod, LMA = locomotor activity in anopen field, GS FL = grip strength of the forelimbs, GS 4L = gripstrength of the fore- and hind limbs, TM CL = swimming T-maze cue learning, TM CR = swimming T-maze cue reversal,FC CS = fear conditioning cue recall (total immobility over 25cue exposures), FC CT = fear conditioning contextual recall,OD = odor discrimination, SI = social interaction.(PDF)

Table S4. Correlations of behavioral measures takenbetween 15 and 19 weeks. Correlations of performance atbehavioral tasks tested between 15 and 19 weeks, presentedas Pearson r values. LMA = locomotor activity in an open field,GS FL = grip strength of the forelimbs, GS 4L = grip strength ofthe fore- and hind limbs, TM CL = swimming T-maze cuelearning, TM CR = swimming T-maze cue reversal.*Statistically significant after Bonferroni Correction (adjusted pvalue 0.005).(PDF)

Table S5. Correlations of MRI measures. Correlations ofmeasures of pathological burden assessed through MRI acrosstime, presented as Pearson r values. STR = striatum, CTX =cortex, HIPP = hippocampus, CC = corpus callosum, WB =whole brain, MUSC = muscle tissue. *Statistically significantafter Bonferroni Correction (adjusted p value 0.0011).(PDF)

Table S6. Correlations of behavioral abnormalities againstbrain abnormalities determined through MRI. Correlationsof behavioral data against age-matched MRI measures takenbetween either 6 and 12 weeks (matched to 9 week MRI data)and 15 and 19 weeks (matched to 17 week MRI data),presented as Pearson r values. RR = rotarod, LMA = locomotoractivity in an open field, GS FL = grip strength of the forelimbs,GS 4L = grip strength of the fore- and hind limbs, TM CL =swimming T-maze cue learning, TM CR = swimming T-mazecue reversal, FC CS = fear conditioning cue recall (totalimmobility over 25 cue exposures), FC CT = fear conditioningcontextual recall, OD = odor discrimination, SI = socialinteraction, STR = striatum, CTX = cortex, HIPP =hippocampus, CC = corpus callosum, WB = whole brain,MUSC = muscle tissue. *Statistically significant after BonferroniCorrection (adjusted p value 0.001).(PDF)

Table S7. Correlations of mHTT abundance acrossdifferent brain regions. Correlations of total detectable mHTT(Total aggregated mHTT) and nuclear inclusions (Nuc mHTT)across the six brain regions investigated in 19 week old R6/1mice only, presented as Pearson r values. STR = striatum,CTX = cortex, DG = dentate gyrus, CA1 = hippocampal CA1subfield, CA2 = hippocampal CA2 subfield, CA3 = hippocampalCA3 subfield. *Statistically significant after BonferroniCorrection (adjusted p value 0.0008).(PDF)

Table S8. Correlations of neuronal characteristics.Correlations of neuronal number (Neur no.), density (Neurdens.) and regional volume determined through stereologicalanalyses of NeuN-strained brain sections in 19 week old mice,presented as Pearson r values. STR = striatum, M1 CTX = M1cortex. *Statistically significant after Bonferroni Correction(adjusted p value 0.0033).(PDF)



Table S9. Correlations of mHTT levels versus MRImeasures of brain abnormalities. Correlations of both post-mortem (at 19 weeks) total aggregated mHTT levels andnuclear inclusions (Nuc mHTT) against measures of brainpathology through MRI at 17 weeks, presented as Pearson rvalues. STR = striatum, CTX = cortex, DG = dentate gyrus,CA1 = hippocampal CA1 subfield, CA2 = hippocampal CA2subfield, CA3 = hippocampal CA3 subfield, HIPP =hippocampus, CC = corpus callosum, WB = whole brain,MUSC = muscle tissue.(PDF)

Table S10. Correlations of neuronal characteristics versusMRI measures of brain abnormalities. Correlations of post-mortem neuronal number (Neur no.), density (Neur dens.) andregional volume determined through stereology on NeuN-stained brain sections against measures of brain pathologythrough MRI at 17 weeks of age, presented as Pearson rvalues. STR = striatum, M1 CTX = primary motor cortex, CTX =cortex, HIPP = hippocampus, CC = corpus callosum, WB =whole brain, MUSC = muscle tissue. *Statistically significantafter Bonferroni Correction (adjusted p value 0.0033).(PDF)

Table S11. Correlations of mHTT levels versus behavioralperformance. Correlations of post-mortem (19 weeks) totalmHTT and nuclear inclusions (Nuc mHTT) against behavioralphenotypes exhibited by R6/1 mice between 15 and 19 weeksof age, presented as Pearson r values. LMA = locomotoractivity in an open field, GS FL = grip strength of the forelimbs,GS 4L = grip strength of the fore- and hind limbs, TM CL =swimming T-maze cue learning, TM CR = swimming T-mazecue reversal, STR = striatum, CTX = cortex, DG = dentategyrus, CA1 = hippocampal CA1 subfield, CA2 = hippocampal

CA2 subfield, CA3 = hippocampal CA3 subfield. *Statisticallysignificant after Bonferroni Correction (adjusted p value0.0017).(PDF)

Table S12. Correlations of neuronal characteristics versusbehavioral performance. Correlations of post-mortem (19weeks) neuronal number (Neur no.), density (Neur dens.) andregional volume determined through stereology on NeuN-stained brain sections against behavioral performancerecorded between 15 and 19 weeks of age, presented asPearson r values. LMA = locomotor activity in an open field, GSFL = grip strength of the forelimbs, GS 4L = grip strength of thefore- and hind limbs, TM CL = swimming T-maze cue learning,TM CR = swimming T-maze cue reversal, STR = striatum, M1CTX = primary motor cortex. *Statistically significant afterBonferroni Correction (adjusted p value 0.0033).(PDF)

Acknowledgements

We thank Dr. Donna Smith for providing R6/1 males forbreeding, Maik Stille for assistance with semi-automated imageanalysis and Prof. Jon Cooper for help with histologicalprocedures.

Author Contributions

Conceived and designed the experiments: IR RG GPB MM.Performed the experiments: IR EJS WRC TAW RG. Analyzedthe data: IR EJS WRC TAW. Contributed reagents/materials/analysis tools: WRC. Wrote the manuscript: IR EJS WRC GPBMM.

References

1. (1993) A novel gene containing a trinucleotide repeat that is expandedand unstable on Huntingtons disease chromosomes. The Huntington'sDisease Collaborative Research. (1993) Group. Cell 72: 971-983 doi:10.1016/0092-8674(93)90585-E. PubMed: 8458085. doi:10.1016/0092-8674(93)90585-E. PubMed: 8458085.

2. Tabrizi SJ, Scahill RI, Durr A, Roos RA, Leavitt BR et al. (2011)Biological and clinical changes in premanifest and early stageHuntington's disease in the TRACK-HD study: the 12-monthlongitudinal analysis. Lancet Neurol 10: 31-42. doi:10.1016/S1474-4422(10)70276-3. PubMed: 21130037.

3. Stout JC, Jones R, Labuschagne I, O'Regan AM, Say MJ et al. (2012)Evaluation of longitudinal 12 and 24 month cognitive outcomes inpremanifest and early Huntington's disease. J Neurol NeurosurgPsychiatry 83: 687-694. doi:10.1136/jnnp-2011-301940. PubMed:22566599.

4. Delmaire C, Dumas EM, Sharman MA, van den Bogaard SJ,Valabregue R et al. (2013) The structural correlates of functionaldeficits in early huntington's disease. Hum Brain Mapp, 34: 2141–53.PubMed: 22438242.

5. Mangiarini L, Sathasivam K, Seller M, Cozens B, Harper A et al. (1996)Exon 1 of the HD gene with an expanded CAG repeat is sufficient tocause a progressive neurological phenotype in transgenic mice. Cell87: 493-506. doi:10.1016/S0092-8674(00)81369-0. PubMed: 8898202.

6. Schilling G, Becher MW, Sharp AH, Jinnah HA, Duan K et al. (1999)Intranuclear inclusions and neuritic aggregates in transgenic miceexpressing a mutant N-terminal fragment of huntingtin. Hum Mol Genet8: 397-407. doi:10.1093/hmg/8.3.397. PubMed: 9949199.

7. Slow EJ, van Raamsdonk J, Rogers D, Coleman SH, Graham RK et al.(2003) Selective striatal neuronal loss in a YAC128 mouse model ofHuntington disease. Hum Mol Genet 12: 1555-1567. doi:10.1093/hmg/ddg169. PubMed: 12812983.

8. Gray M, Shirasaki DI, Cepeda C, André VM, Wilburn B et al. (2008)Full-length human mutant huntingtin with a stable polyglutamine repeatcan elicit progressive and selective neuropathogenesis in BACHDmice. J Neurosci 28: 6182-6195. doi:10.1523/JNEUROSCI.0857-08.2008. PubMed: 18550760.

9. Lin CH, Tallaksen-Greene S, Chien WM, Cearley JA, Jackson WS etal. (2001) Neurological abnormalities in a knock-in mouse model ofHuntington's disease. Hum Mol Genet 10: 137-144. doi:10.1093/hmg/10.2.137. PubMed: 11152661.

10. Wheeler VC, White JK, Gutekunst CA, Vrbanac V, Weaver M et al.(2000) Long glutamine tracts cause nuclear localization of a novel formof huntingtin in medium spiny striatal neurons in HdhQ92 and HdhQ111knock-in mice. Hum Mol Genet 9: 503-513. doi:10.1093/hmg/9.4.503.PubMed: 10699173.

11. Menalled LB, Sison JD, Wu Y, Olivieri M, Li XJ et al. (2002) Early motordysfunction and striosomal distribution of huntingtin microaggregates inHuntington's disease knock-in mice. J Neurosci 22: 8266-8276.PubMed: 12223581.

12. Woodman B, Butler R, Landles C, Lupton MK, Tse J et al. (2007) TheHdh(Q150/Q150) knock-in mouse model of HD and the R6/2 exon 1model develop comparable and widespread molecular phenotypes.Brain. Res Bull 72: 83-97. doi:10.1016/j.brainresbull.2006.11.004.

13. Kuhn A, Goldstein DR, Hodges A, Strand AD, Sengstag T et al. (2007)Mutant huntingtin's effects on striatal gene expression in mice



http://dx.doi.org/10.1016/S1474-4422(10)70276-3

http://dx.doi.org/10.1016/S1474-4422(10)70276-3

http://www.ncbi.nlm.nih.gov/pubmed/21130037

http://dx.doi.org/10.1136/jnnp-2011-301940



http://dx.doi.org/10.1016/S0092-8674(00)81369-0


http://dx.doi.org/10.1093/hmg/8.3.397


http://dx.doi.org/10.1093/hmg/ddg169

http://dx.doi.org/10.1093/hmg/ddg169


http://dx.doi.org/10.1523/JNEUROSCI.0857-08.2008









http://dx.doi.org/10.1016/j.brainresbull.2006.11.004

recapitulate changes observed in human Huntington's disease brainand do not differ with mutant huntingtin length or wild-type huntingtindosage. Hum Mol Genet 16: 1845-1861. doi:10.1093/hmg/ddm133.PubMed: 17519223.

14. Moffitt H, McPhail GD, Woodman B, Hobbs C, Bates GP (2009)Formation of polyglutamine inclusions in a wide range of non-CNStissues in the HdhQ150 knock-in mouse model of Huntington's disease.PLOS ONE 4: e8025. doi:10.1371/journal.pone.0008025. PubMed:19956633.

15. Sathasivam K, Lane A, Legleiter J, Warley A, Woodman B et al. (2010)Identical oligomeric and fibrillar structures captured from the brains ofR6/2 and knock-in mouse models of Huntington's disease. Hum MolGenet 19: 65-78. doi:10.1093/hmg/ddq162. PubMed: 19825844.

16. Labbadia J, Cunliffe H, Weiss A, Katsyuba E, Sathasivam K et al.(2011) Altered chromatin architecture underlies progressive impairmentof the heat shock response in mouse models of Huntington disease. JClin Invest 121: 3306-3319. doi:10.1172/JCI57413. PubMed:21785217.

17. Landles C, Sathasivam K, Weiss A, Woodman B, Moffitt H et al. (2010)Proteolysis of mutant huntingtin produces an exon 1 fragment thataccumulates as an aggregated protein in neuronal nuclei in Huntingtondisease. J Biol Chem 285: 8808-8823. doi:10.1074/jbc.M109.075028.PubMed: 20086007.

18. Sathasivam K, Neueder A, Gipson TA, Landles C, Benjamin AC et al.(2013) Aberrant splicing of HTT generates the pathogenic exon 1protein in Huntington disease. Proc Natl Acad Sci U S A 110:2366-2370. doi:10.1073/pnas.1221891110. PubMed: 23341618.

19. Rattray I, Smith E, Gale R, Matsumoto K, Bates GP et al. (2013)Correlations of Behavioral Deficits with Brain Pathology Assessedthrough Longitudinal MRI and Histopathology in the R6/2 Mouse Modelof HD. PLOS ONE 8: e60012. doi:10.1371/journal.pone.0060012.PubMed: 23593159.

20. Brooks SP, Janghra N, Workman VL, Bayram-Weston Z, Jones L et al.(2012) Longitudinal analysis of the behavioural phenotype in R6/1(C57BL/6J) Huntington's disease transgenic mice. Brain. Res Bull 88:94-103. doi:10.1016/j.brainresbull.2011.01.010.

21. Bayram-Weston Z, Jones L, Dunnett SB, Brooks SP (2012) Light andelectron microscopic characterization of the evolution of cellularpathology in the R6/1 Huntington's disease transgenic mice. Brain. ResBull 88: 104-112. doi:10.1016/j.brainresbull.2011.07.009.

22. Abada YS, Schreiber R, Ellenbroek B (2013) Motor, emotional andcognitive deficits in adult BACHD mice: a model for Huntington'sdisease. Behav Brain Res 238: 243-251. doi:10.1016/j.bbr.2012.10.039. PubMed: 23123142.

23. Brooks S, Higgs G, Janghra N, Jones L, Dunnett SB (2012)Longitudinal analysis of the behavioural phenotype in YAC128(C57BL/6J) Huntington's disease transgenic mice. Brain. Res Bull 88:113-120. doi:10.1016/j.brainresbull.2010.05.005.

24. Brooks SP, Janghra N, Higgs GV, Bayram-Weston Z, Heuer A et al.(2012) Selective cognitive impairment in the YAC128 Huntington'sdisease mouse. Brain. Res Bull 88: 121-129. doi:10.1016/j.brainresbull.2011.05.010.

25. Brooks SP, Jones L, Dunnett SB (2012) Longitudinal analyses ofoperant performance on the serial implicit learning task (SILT) in theYAC128 Huntington's disease mouse line. Brain. Res Bull 88: 130-136.doi:10.1016/j.brainresbull.2011.06.008.

26. Bayram-Weston Z, Jones L, Dunnett SB, Brooks SP (2012) Light andelectron microscopic characterization of the evolution of cellularpathology in YAC128 Huntington's disease transgenic mice. Brain. ResBull 88: 137-147. doi:10.1016/j.brainresbull.2011.05.005.

27. Brooks S, Higgs G, Jones L, Dunnett SB (2012) Longitudinal analysisof the behavioural phenotype in HdhQ92 Huntington's disease knock-inmice. Brain. Res Bull 88: 148-155. doi:10.1016/j.brainresbull.2010.05.003.

28. Trueman RC, Dunnett SB, Jones L, Brooks SP (2012) Five choiceserial reaction time performance in the HdhQ92 mouse model ofHuntington's disease. Brain. Res Bull 88: 163-170. doi:10.1016/j.brainresbull.2011.10.019.

29. Bayram-Weston Z, Jones L, Dunnett SB, Brooks SP (2012) Light andelectron microscopic characterization of the evolution of cellularpathology in HdhQ92 Huntington's disease knock-in mice. Brain. ResBull 88: 171-181. doi:10.1016/j.brainresbull.2011.03.013.

30. Brooks S, Higgs G, Jones L, Dunnett SB (2012) Longitudinal analysisof the behavioural phenotype in Hdh(CAG)150 Huntington's diseaseknock-in mice. Brain. Res Bull 88: 182-188. doi:10.1016/j.brainresbull.2010.05.004.

31. Trueman RC, Jones L, Dunnett SB, Brooks SP (2012) Early onsetdeficits on the delayed alternation task in the Hdh(Q92) knock-in mouse

model of Huntington's disease. Brain. Res Bull 88: 156-162. doi:10.1016/j.brainresbull.2011.03.012.

32. Bayram-Weston Z, Torres EM, Jones L, Dunnett SB, Brooks SP (2012)Light and electron microscopic characterization of the evolution ofcellular pathology in the Hdh(CAG)150 Huntington's disease knock-inmouse. Brain. Res Bull 88: 189-198. doi:10.1016/j.brainresbull.2011.03.014.

33. Giles P, Elliston L, Higgs GV, Brooks SP, Dunnett SB et al. (2012)Longitudinal analysis of gene expression and behaviour in theHdhQ150 mouse model of Huntington's disease. Brain. Res Bull 88:199-209. doi:10.1016/j.brainresbull.2011.10.001.

34. Heikkinen T, Lehtimäki K, Vartiainen N, Puoliväli J, Hendricks SJ et al.(2012) Characterization of neurophysiological and behavioral changes,MRI brain volumetry and 1H MRS in zQ175 knock-in mouse model ofHuntington's disease. PLOS ONE 7: e50717. doi:10.1371/journal.pone.0050717. PubMed: 23284644.

35. Menalled LB, Kudwa AE, Miller S, Fitzpatrick J, Watson-Johnson J etal. (2012) Comprehensive behavioral and molecular characterization ofa new knock-in mouse model of Huntington's disease: zQ175. PLOSONE 7: e49838. doi:10.1371/journal.pone.0049838. PubMed:23284626.

36. Izquierdo A, Wellman CL, Holmes A (2006) Brief uncontrollable stresscauses dendritic retraction in infralimbic cortex and resistance to fearextinction in mice. J Neurosci 26: 5733-5738. doi:10.1523/JNEUROSCI.0474-06.2006. PubMed: 16723530.

37. Macbeth AH, Edds JS, Young WS 3rd (2009) Housing conditions andstimulus females: a robust social discrimination task for studying malerodent social recognition. Nat Protoc 4: 1574-1581. doi:10.1038/nprot.2009.141. PubMed: 19816420.

38. El-Akabawy G, Rattray I, Johansson SM, Gale R, Bates G et al. (2012)Implantation of undifferentiated and pre-differentiated human neuralstem cells in the R6/2 transgenic mouse model of Huntington's disease.BMC Neurosci 13: 97. doi:10.1186/1471-2202-13-97. PubMed:22876937.

39. Crum WR, Giampietro VP, Smith EJ, Gorenkova N, Stroemer RP et al.(2013) A comparison of automated anatomical-behavioural mappingmethods in a rodent model of stroke. J Neurosci Methods 218:170-183. doi:10.1016/j.jneumeth.2013.05.009. PubMed: 23727124.

40. Crum WR, Modo M, Vernon AC, Barker GJ, Williams SC (2013)Registration of challenging pre-clinical brain images. J NeurosciMethods 216: 62-77. doi:10.1016/j.jneumeth.2013.03.015. PubMed:23558335.

41. Crum WR, Tanner C, Hawkes DJ (2005) Anisotropic multi-scale fluidregistration: evaluation in magnetic resonance breast imaging. PhysMed Biol 50: 5153-5174. doi:10.1088/0031-9155/50/21/014. PubMed:16237247.

42. Freeborough PA, Fox NC (1998) Modeling brain deformations inAlzheimer disease by fluid registration of serial 3D MR images. JComput Assist Tomogr 22: 838-843. doi:10.1097/00004728-199809000-00031. PubMed: 9754126.

43. Fox NC, Crum WR, Scahill RI, Stevens JM, Janssen JC et al. (2001)Imaging of onset and progression of Alzheimer's disease with voxel-compression mapping of serial magnetic resonance images. Lancet358: 201-205. doi:10.1016/S0140-6736(01)05408-3. PubMed:11476837.

44. Sathasivam K, Woodman B, Mahal A, Bertaux F, Wanker EE et al.(2001) Centrosome disorganization in fibroblast cultures derived fromR6/2 Huntington's disease (HD) transgenic mice and HD patients. HumMol Genet 10: 2425-2435. doi:10.1093/hmg/10.21.2425. PubMed:11689489.

45. Mouton PR (2011) Unbiased stereology: A consise guide. Baltimore,MD: John Hopkins University Press.

46. Sadikot AF, Sasseville R (1997) Neurogenesis in the mammalianneostriatum and nucleus accumbens: parvalbumin-immunoreactiveGABAergic interneurons. J Comp Neurol 389: 193-211. doi:10.1002/(SICI)1096-9861(19971215)389:2. PubMed: 9416916.

47. Paxinos G, Franklin KBJ (2001) The mouse brain in stereotaxiccoordinates. San Diego, CA: Academic Press.

48. Genovese CR, Lazar NA, Nichols T (2002) Thresholding of statisticalmaps in functional neuroimaging using the false discovery rate.NeuroImage 15: 870-878. doi:10.1006/nimg.2001.1037. PubMed:11906227.

49. Hickey MA, Gallant K, Gross GG, Levine MS, Chesselet MF (2005)Early behavioral deficits in R6/2 mice suitable for use in preclinical drugtesting. Neurobiol Dis 20: 1-11. doi:10.1016/j.nbd.2005.01.024.PubMed: 16137562.

50. Stack EC, Kubilus JK, Smith K, Cormier K, Del Signore SJ et al. (2005)Chronology of behavioral symptoms and neuropathological sequela in



http://dx.doi.org/10.1093/hmg/ddm133


http://dx.doi.org/10.1371/journal.pone.0008025


http://dx.doi.org/10.1093/hmg/ddq162


http://dx.doi.org/10.1172/JCI57413


http://dx.doi.org/10.1074/jbc.M109.075028


http://dx.doi.org/10.1073/pnas.1221891110






http://dx.doi.org/10.1016/j.bbr.2012.10.039

http://dx.doi.org/10.1016/j.bbr.2012.10.039


























http://dx.doi.org/10.1038/nprot.2009.141

http://dx.doi.org/10.1038/nprot.2009.141


http://dx.doi.org/10.1186/1471-2202-13-97


http://dx.doi.org/10.1016/j.jneumeth.2013.05.009


http://dx.doi.org/10.1016/j.jneumeth.2013.03.015


http://dx.doi.org/10.1088/0031-9155/50/21/014


http://dx.doi.org/10.1097/00004728-199809000-00031


http://dx.doi.org/10.1016/S0140-6736(01)05408-3




http://dx.doi.org/10.1002/(SICI)1096-9861(19971215)389:2

http://dx.doi.org/10.1002/(SICI)1096-9861(19971215)389:2


http://dx.doi.org/10.1006/nimg.2001.1037


http://dx.doi.org/10.1016/j.nbd.2005.01.024


R6/2 Huntington's disease transgenic mice. J Comp Neurol 490:354-370. doi:10.1002/cne.20680. PubMed: 16127709.

51. Smith KM, Matson S, Matson WR, Cormier K, Del Signore SJ et al.(2006) Dose ranging and efficacy study of high-dose coenzyme Q10formulations in Huntington's disease mice. Biochim Biophys Acta 1762:616-626. doi:10.1016/j.bbadis.2006.03.004. PubMed: 16647250.

52. Mazarakis NK, Cybulska-Klosowicz A, Grote H, Pang T, Van Dellen Aet al. (2005) Deficits in experience-dependent cortical plasticity andsensory-discrimination learning in presymptomatic Huntington'sdisease mice. J Neurosci 25: 3059-3066. doi:10.1523/JNEUROSCI.4320-04.2005. PubMed: 15788762.

53. Naver B, Stub C, Møller M, Fenger K, Hansen AK et al. (2003)Molecular and behavioral analysis of the R6/1 Huntington's diseasetransgenic mouse. Neuroscience 122: 1049-1057. doi:10.1016/j.neuroscience.2003.08.053. PubMed: 14643771.

54. Bolivar VJ, Manley K, Messer A (2004) Early exploratory behaviorabnormalities in R6/1 Huntington's disease transgenic mice. Brain Res1005: 29-35. doi:10.1016/j.brainres.2004.01.021. PubMed: 15044061.

55. Hodges A, Hughes G, Brooks S, Elliston L, Holmans P et al. (2008)Brain gene expression correlates with changes in behavior in the R6/1mouse model of Huntington's disease. Genes Brain Behav 7: 288-299.doi:10.1111/j.1601-183X.2007.00350.x. PubMed: 17696994.

56. Bissonette GB, Martins GJ, Franz TM, Harper ES, Schoenbaum G etal. (2008) Double dissociation of the effects of medial and orbitalprefrontal cortical lesions on attentional and affective shifts in mice. JNeurosci 28: 11124-11130. doi:10.1523/JNEUROSCI.2820-08.2008.PubMed: 18971455.

57. Bissonette GB, Powell EM (2012) Reversal learning and attentional set-shifting in mice. Neuropharmacology 62: 1168-1174. doi:10.1016/j.neuropharm.2011.03.011. PubMed: 21439304.

58. Miller BR, Walker AG, Barton SJ, Rebec GV (2011) DysregulatedNeuronal Activity Patterns Implicate Corticostriatal Circuit Dysfunctionin Multiple Rodent Models of Huntington's Disease. Frontiers inSystems Neuroscience 5: 26. PubMed: 21629717.

59. Rebec GV, Conroy SK, Barton SJ (2006) Hyperactive striatal neuronsin symptomatic Huntington R6/2 mice: variations with behavioral stateand repeated ascorbate treatment. Neuroscience 137: 327-336. doi:10.1016/j.neuroscience.2005.08.062. PubMed: 16257492.

60. Walker AG, Miller BR, Fritsch JN, Barton SJ, Rebec GV (2008) Alteredinformation processing in the prefrontal cortex of Huntington's diseasemouse models. J Neurosci 28: 8973-8982. doi:10.1523/JNEUROSCI.2804-08.2008. PubMed: 18768691.

61. Ruocco HH, Bonilha L, Li LM, Lopes-Cendes I, Cendes F (2008)Longitudinal analysis of regional grey matter loss in Huntingtondisease: effects of the length of the expanded CAG repeat. J NeurolNeurosurg Psychiatry 79: 130-135. doi:10.1136/jnnp.2007.116244.PubMed: 17615168.

62. Tabrizi SJ, Scahill RI, Owen G, Durr A, Leavitt BR et al. (2013)Predictors of phenotypic progression and disease onset in premanifestand early-stage Huntington's disease in the TRACK-HD study: analysisof 36-month observational data. Lancet Neurol 12: 637-649. doi:10.1016/S1474-4422(13)70088-7. PubMed: 23664844.

63. Dominguez DJ, Egan GF, Gray MA, Poudel GR, Churchyard A et al.(2013) Multi-Modal Neuroimaging in Premanifest and EarlyHuntington's Disease: 18 Month Longitudinal Data from the IMAGE-HD

Study. PLOS ONE 8: e74131. doi:10.1371/journal.pone.0074131.PubMed: 24066104.

64. Aylward EH, Liu D, Nopoulos PC, Ross CA, Pierson RK et al. (2012)Striatal volume contributes to the prediction of onset of Huntingtondisease in incident cases. Biol Psychiatry 71: 822-828. doi:10.1016/j.biopsych.2011.07.030. PubMed: 21907324.

65. Crum WR, Giampietro VP, Smith EJ, Gorenkova N, Stroemer RP et al.(2013) A comparison of automated anatomical-behavioural mappingmethods in a rodent model of stroke. J Neuroscience Methods.

66. Vernon AC, Crum WR, Johansson SM, Modo M (2011) Evolution ofextra-nigral damage predicts behavioural deficits in a rat proteasomeinhibitor model of Parkinson's disease. PLOS ONE 6: e17269. doi:10.1371/journal.pone.0017269. PubMed: 21364887.

67. Ma L, Morton AJ, Nicholson LF (2003) Microglia density decreases withage in a mouse model of Huntington's disease. Glia 43: 274-280. doi:10.1002/glia.10261. PubMed: 12898706.

68. Kusakabe M, Mangiarini L, Laywell ED, Bates GP, Yoshiki A et al.(2001) Loss of cortical and thalamic neuronal tenascin-C expression ina transgenic mouse expressing exon 1 of the human Huntingtondisease gene. J Comp Neurol 430: 485-500. doi:10.1002/1096-9861(20010219)430:4. PubMed: 11169482.

69. Hong SL, Cossyleon D, Hussain WA, Walker LJ, Barton SJ et al.(2012) Dysfunctional behavioral modulation of corticostriatalcommunication in the R6/2 mouse model of Huntington's disease.PLOS ONE 7: e47026. doi:10.1371/journal.pone.0047026. PubMed:23056565.

70. Hong SL, Barton SJ, Rebec GV (2012) Altered neural and behavioraldynamics in Huntington's disease: an entropy conservation approach.PLOS ONE 7: e30879. doi:10.1371/journal.pone.0030879. PubMed:22292068.

71. Miller BR, Walker AG, Barton SJ, Rebec GV (2011) DysregulatedNeuronal Activity Patterns Implicate Corticostriatal Circuit Dysfunctionin Multiple Rodent Models of Huntington's Disease. Front Syst Neurosci5: 26. PubMed: 21629717.

72. Maier SF (1984) Learned helplessness and animal models ofdepression. Prog Neuropsychopharmacol Biol Psychiatry 8: 435-446.doi:10.1016/0278-5846(84)90120-9. PubMed: 6385140.

73. Olsson IAS, Nevison CM, Patterson-Kane EG, Sherwin CM, Van deWeerd HA et al. (2003) Understanding behaviour: the relevance ofethological approaches in laboratory animal science. Applied AnimalBehaviour Science 81: 245-264. doi:10.1016/S0168-1591(02)00285-X.

74. de Kloet ER, Oitzl MS, Joëls M (1999) Stress and cognition: arecorticosteroids good or bad guys? Trends Neurosci 22: 422-426. doi:10.1016/S0166-2236(99)01438-1. PubMed: 10481183.

75. Pardon MC, Rattray I (2008) What do we know about the long-termconsequences of stress on ageing and the progression of age-relatedneurodegenerative disorders? Neurosci Biobehav Rev 32: 1103-1120.doi:10.1016/j.neubiorev.2008.03.005. PubMed: 18436304.

76. Björkqvist M, Petersén A, Bacos K, Isaacs J, Norlén P et al. (2006)Progressive alterations in the hypothalamic-pituitary-adrenal axis in theR6/2 transgenic mouse model of Huntington's disease. Hum Mol Genet15: 1713-1721. doi:10.1093/hmg/ddl094. PubMed: 16613897.

77. Du X, Leang L, Mustafa T, Renoir T, Pang TY et al. (2012)Environmental enrichment rescues female-specific hyperactivity of thehypothalamic-pituitary-adrenal axis in a model of Huntington's disease.Transl Psychiatry 2: e133. doi:10.1038/tp.2012.58. PubMed: 22760557.



http://dx.doi.org/10.1002/cne.20680


http://dx.doi.org/10.1016/j.bbadis.2006.03.004





http://dx.doi.org/10.1016/j.neuroscience.2003.08.053



http://dx.doi.org/10.1016/j.brainres.2004.01.021


http://dx.doi.org/10.1111/j.1601-183X.2007.00350.x




http://dx.doi.org/10.1016/j.neuropharm.2011.03.011

http://dx.doi.org/10.1016/j.neuropharm.2011.03.011








http://dx.doi.org/10.1136/jnnp.2007.116244


http://dx.doi.org/10.1016/S1474-4422(13)70088-7




http://dx.doi.org/10.1016/j.biopsych.2011.07.030

http://dx.doi.org/10.1016/j.biopsych.2011.07.030




http://dx.doi.org/10.1002/glia.10261


http://dx.doi.org/10.1002/1096-9861(20010219)430:4







http://dx.doi.org/10.1016/0278-5846(84)90120-9


http://dx.doi.org/10.1016/S0168-1591(02)00285-X

http://dx.doi.org/10.1016/S0166-2236(99)01438-1


http://dx.doi.org/10.1016/j.neubiorev.2008.03.005


http://dx.doi.org/10.1093/hmg/ddl094


http://dx.doi.org/10.1038/tp.2012.58


assessed through longitudinal mri and histopathology *,1,4...

Documents